Comparison of Outcomes Between BRCA Pathogenic Variant Carriers Undergoing Breast-Conserving Surgery Versus Mastectomy

Abstract

Introduction:

Although outcomes are similar following breast-conserving surgery (BCS) or mastectomy among sporadic breast cancer patients, data are mixed for women with a germline BRCA mutation. We sought to compare outcomes among a modern cohort of BRCA mutation carriers undergoing BCS versus mastectomy.

Methods:

Women with a BRCA mutation and an index breast cancer from 2006–2015 were retrospectively identified from institutional databases. Factors including date of genetic testing, clinicopathologic details, and treatment characteristics were identified. Subsequent locoregional recurrence (LRR), distant recurrence, contralateral breast cancer (CBC), breast cancer-specific survival (BCSS), and overall survival (OS) events were compared between groups.

Results:

395 BRCA mutation carriers with 424 cancers were identified. Surgical treatment included BCS for 99 cancers, and mastectomy for 325 cancers. Patients choosing mastectomy were more likely to have bilateral breast cancer, be younger/premenopausal, and be aware of their genetic status before surgery, and less likely to receive radiation therapy (p<0.001). At 7.9 years median follow-up, LRR, distant recurrence, BCSS, and OS rates did not differ between groups. CBC occurred in 5 versus 0 women treated with unilateral versus bilateral surgery, respectively, resulting is a 10-year estimated CBC risk of 14% among unilateral breast surgery patients (p<0.001).

Conclusions:

With nearly 8 years follow-up, we report no difference in LRR, BCSS, and OS among BRCA mutation carriers who underwent BCS or mastectomy; however, we report a higher incidence of CBC among those undergoing unilateral breast surgery. These data support BCS as an option for BRCA mutation carriers willing to continue high-risk screening.

INTRODUCTION

Women with a BRCA1 or BRCA2 germline mutation are at a substantially higher risk for the development of breast cancer compared to the general population, with a cumulative lifetime risk of breast cancer of approximately 70%.^1–3 Additionally, this population is predisposed to the development of a metachronous contralateral breast cancer. Patients with BRCA1 mutations have a 40% cumulative risk, and those with BRCA2 mutation have a 26% cumulative risk, of having a second event in their previously unaffected breast in the 20 years following an index breast cancer diagnosis.²

Among women with sporadic breast cancer, it is well established that oncologic outcomes, including recurrence and survival, are similar following breast-conserving surgery (BCS) versus mastectomy.⁴ While women with a BRCA mutation often choose mastectomy for both therapeutic and prophylactic intent to prevent a second breast cancer event, the literature assessing outcomes between breast conservation and mastectomy among this population has been limited and somewhat mixed.^5–8 The potential oncologic benefit of mastectomy may differ among women with a sporadic breast cancer compared to those with a germline BRCA mutation, given their underlying predisposition to breast cancer development.

Here we report on a contemporary cohort of BRCA mutation carriers who opted for BCS versus mastectomy in the treatment of an index breast cancer presentation, and report on locoregional recurrence, contralateral breast cancer, distant recurrence, and survival outcomes.

METHODS

Data Source

Following Memorial Sloan Kettering Cancer Center institutional review board approval, we identified women presenting with an index breast cancer diagnosis (either ductal carcinoma in situ [DCIS] or invasive breast cancer) from 2006–2015 who were either known or subsequently identified to have a BRCA1 or BRCA2 pathogenic variant. We also included women with a BRCA mutation who were incidentally found to have breast cancer at the time of prophylactic risk-reducing bilateral mastectomy. Women with variants of uncertain significance for BRCA1 and BRCA2, men, those with a prior DCIS or invasive breast cancer diagnosis, those with a phyllodes tumor, or those with other cancer diagnoses at the time of their breast cancer diagnosis (ovarian, pancreatic) were excluded.

Variables

Patient characteristics, including age at diagnosis, menopausal status, date of genetic testing, and genetic testing results were collected at the time of breast cancer diagnosis. Breast cancer diagnostic and pathologic characteristics, including method of cancer detection; tumor size; tumor histology; nodal status; and estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 (HER2) status were captured. Both surgeon assessment of BCS eligibility, and the final surgical treatment for each patient, were identified and captured. The final surgical treatment was classified as BCS, unilateral mastectomy, or bilateral mastectomy. Axillary surgery details were recorded as no axillary surgery, sentinel lymph node biopsy (SLNB) alone, or axillary lymph node dissection (ALND). If a patient initially opted for breast conservation, but subsequently underwent bilateral mastectomy within 6 months of their index surgical date, they were categorized within the bilateral mastectomy cohort for analysis. Women who underwent initial BCS and elected to undergo bilateral mastectomy more than 6 months from their index surgical treatment were categorized as having had breast conservation and censored at the date of bilateral mastectomy for contralateral breast event analysis. Additional treatment received, including neoadjuvant or adjuvant chemotherapy, hormone therapy, and radiation therapy were recorded.

All subsequent breast cancer recurrence events were collected, including locoregional recurrence (LRR), defined as recurrence in the ipsilateral breast or chest wall or ipsilateral axillary, supraclavicular, or internal mammary nodes; distant recurrence; contralateral breast cancer; and breast cancer-specific survival (BCSS) and overall survival (OS), during the follow-up period.

Statistical Analysis

To assess the characteristics (including type of surgery) associated with the incidence of recurrence, incidence of death due to any cause, breast cancer-specific death, and the incidence of contralateral breast cancer in patients with a BRCA1 or BRCA2 mutation, we performed a univariate analysis (UVA) comparing characteristics between those individuals who developed an event and those who did not develop an event. Continuous covariates were compared using the t-test or Wilcoxon rank sum test, and categorical covariates were compared using Fisher’s exact test or the chi-squared test. Based on the results of the UVA, a multivariable Cox regression analysis (MVA) was then conducted by selecting covariates from the UVA, which were significant at a type I error rate (α) of 0.05. The final list of covariates in the MVA were chosen using the backward elimination procedure. Since we are making multiple comparisons by considering different outcomes, we performed a Bonferroni adjustment in the MVA and thereby adjusted the type I error rate to 0.0125 (α*). Kaplan-Meier curves were constructed to compare survival between those patients who underwent BCS and those who underwent mastectomy. Additionally, a comparison of clinicopathological characteristics between the 2 surgical groups is also provided. All statistical analyses were made using R 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Population

395 patients were identified who presented with a diagnosis of DCIS or stage I-III invasive breast cancer between 2006–2015 who were known or found to carry a BRCA1 or BRCA2 pathologic variant. This included 29 patients with bilateral breast cancers at their time of diagnosis for a total of 424 individual breast cancers. The median age at breast cancer diagnosis was 44 years. In total, 209 women had a BRCA1 pathogenic variant and 215 had a BRCA2 pathogenic variant. Median follow up was 7.9 years (Fig. 1).

Fig. 1.

Study population.

BCS breast-conserving surgery, CPM contralateral prophylactic mastectomy

Surgical Details

325 cancers were treated with mastectomy, including 312 treated with bilateral mastectomy and 12 with unilateral mastectomy. The majority of mastectomies were skin-sparing procedures, and 9% (n = 30) were nipple-sparing mastectomies. 99 cancers were treated with BCS, with 43 patients opting to undergo bilateral mastectomy more than 6 months after their index cancer surgery date, with a median time to mastectomy of 19 months (range 7 to 90 months). There were no significant differences in rates of BCS or mastectomy between women with a BRCA1 or BRCA 2 mutation (Table 1). Of those who opted for mastectomy, 77% were considered eligible for BCS at diagnosis. Women who underwent mastectomy were younger than those opting for BCS, age 43 years compared with 48 years, respectively (p < 0.001). Women undergoing mastectomy were more likely to be aware of their BRCA mutation carrier status at the time of surgery, were more likely to be premenopausal, and more likely to present with synchronous bilateral breast cancer compared to women undergoing BCS (all p < 0.001). No significant differences in tumor histology, tumor size, nodal status, or receptor profile were seen between women undergoing BCS versus mastectomy. Women undergoing mastectomy were more likely to undergo axillary staging (BCS 88% versus mastectomy 96%, p = 0.002); however, there was no difference in type of axillary surgery between the surgical cohorts, with 36% and 36% of women undergoing BCS and mastectomy, respectively, receiving ALND.

TABLE 1.

Cohort characteristics

Characteristic	BCS (n = 99)	Mastectomy (n = 325)	p-value
BRCA1	43 (43%)	166 (51%)	0.2
BRCA2	56 (57%)	159 (49%)
BRCA status known at surgery	34 (35%)	258 (80%)	< 0.001
Age at surgery, median (IQR)	48 (42, 58)	43 (38, 50)	< 0.001
Premenopausal	52 (53%)	231 (71%)	< 0.001
Synchronous bilateral breast cancer	3 (3%)	55 (17%)	< 0.001
pT stage
Tis	17 (17%)	67 (21%)	0.5
T1	60 (61%)	176 (54%)
T2–3	22 (22%)	82 (25%)
pN stage
N0	66 (67%)	210 (65%)	> 0.9
N1	24 (24%)	82 (25%)
N2–3	9 (9%)	33 (10%)
ER positive	56 (62%)	164 (50%)	0.5
PR positive	46 (52%)	139 (43%)	0.8
HER2 amplified	7 (8%)	23 (7%)	0.4
Eligible for BCS	99 (100%)	250 (77%)	< 0.001
Axillary surgery	88 (88%)	313 (96%)	0.002
ALND	31 (36%)	118 (36%)	0.8
SLNB	56 (64%)	196 (60%)
Neoadjuvant chemotherapy	9 (9%)	39 (12%)	0.5
Adjuvant chemotherapy	72 (73%)	243 (75%)	0.8
Radiation therapy	64 (65%)	87 (27%)	< 0.001
Endocrine therapy	46 (46%)	186 (57%)	0.1

BCS breast-conserving surgery, IQR interquartile range, ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor 2, ALND axillary lymph node dissection, SLNB sentinel lymph node biopsy

Adjuvant Treatment Details

Overall, 11% of women received neoadjuvant chemotherapy, 74% received adjuvant chemotherapy, and 54% received endocrine therapy. There was no difference in receipt of neoadjuvant or adjuvant chemotherapy and endocrine therapy between the mastectomy and breast conservation cohorts. Patients undergoing BCS were more likely to receive radiation therapy (BCS 65% versus mastectomy 27%, p < 0.001).

Recurrence Events

Overall, there were 67 recurrence events, including 23 LRRs, 44 distant recurrences, and 5 contralateral breast cancer events. From the entire cohort, 40 women died during follow-up, including 36 who died from breast cancer. There were no differences in crude locoregional or distant breast cancer recurrence events between the breast conservation and mastectomy cohorts. Additionally, there was no difference in crude rate of death between the surgical cohorts. Patients opting for unilateral breast surgery were more likely to experience a contralateral breast cancer event (unilateral breast surgery 5% versus bilateral breast surgery 0%, p = 0.001)(Table 2).

TABLE 2.

Crude recurrence events

Recurrence Type	BCS (n = 99)	Mastectomy (n = 325)	p-value
Locoregional	6 (6%)	17 (5%)	> 0.9
Distant	9 (9%)	35 (11%)	0.8
Death	9 (9%)	31 (10%)	0.9
Contralateral breast cancer*	5 (5%)	0	0.001

^*Contralateral breast cancer recurrence among women without synchronous bilateral breast cancer

BCS breast-conserving surgery

Of the 23 LRR events, there were 5 isolated in-breast tumor recurrences in the BCS population, 9 isolated chest wall recurrences following mastectomy, 3 synchronous chest wall and regional recurrences, 3 isolated regional recurrences, and 3 mixed local and distant recurrences. Among 4 patients with isolated ipsilateral breast tumor recurrence (IBTR) and with recurrence details available, 3 appeared to be possible second primary tumors given the recurrence location, in a separate quadrant from the index cancer, while 1 appeared to be a true local recurrence, occurring in the same quadrant with identical tumor markers as the index tumor.

Locoregional and Contralateral Breast Cancer Recurrence Rates

Kaplan-Meier estimates for 5- and 10- year LRR were 7.3% and 7.3%, respectively, for the BCS group, and 3.7% and 6.3%, respectively, for the mastectomy cohort (p = 0.4). Five- and 10-year contralateral breast cancer estimates following unilateral surgery were 3.6% and 13.9%, respectively, compared to 0% and 0%, respectively, following bilateral breast surgery (p < 0.001)(Fig. 2).

Fig. 2.

Locoregional and contralateral breast cancer recurrence rates.

Survival Outcomes

BCSS and OS did not differ by surgical procedure, with Kaplan-Meier 5- and 10- year BCSS rates of 93.6% and 92.2%, respectively, following BCS, and 94.7% and 88.8%, respectively, following mastectomy (p = 0.6). Similarly, OS at 5 years and 10 years following BCS was 92.5% and 88.8%, respectively, compared with 93.3% and 87.5%, respectively, following mastectomy (p = 0.7).

Univariate Analysis

On UVA, assessing factors associated with LRR, only receipt of a contralateral prophylactic mastectomy, and known genetic mutation status at the time of surgery, were significantly associated with a reduced risk of LRR, while having a breast at risk and tumor grade were associated with CBC risk. Younger age, presentation with physical findings, presence of lymphovascular invasion, multifocality, ineligibility for BCS, increasing pathologic T and N stage, and receipt of ALND, adjuvant or neoadjuvant chemotherapy, and radiation therapy were all associated with worse BCSS. Similarly, younger age, presentation with physical findings, ineligibility for BCS, increasing pathologic T and N stage, multifocality, and receipt of ALND, neoadjuvant chemotherapy, and radiation therapy were factors significantly associated with worse OS. Type of surgery was not associated with LRR, BCSS, or OS on UVA.

MVA

On MVA, both having unknown genetic mutation status at the time of surgery (hazard ratio 3.39, 95% confidence interval [CI] 1.3–8.8, p = 0.1) and having undergone a contralateral prophylactic mastectomy (hazard ratio 0.21, 95% CI 0.05–0.95, p = 0.04) were associated with LRR. The interaction between contralateral prophylactic mastectomy and index breast surgery was not significant (p > 0.9). No variables were significant for contralateral breast cancer on MVA. Presentation with physical findings compared to radiologic findings was correlated with worse overall survival (hazard ratio 2.72, 95% CI 1.2–6.2, p = 0.02), as was receipt of neoadjuvant chemotherapy (hazard ratio 2.61, 95% CI 1.3–5.3, p = 0.008), while BCS eligibility was associated with an improved OS (hazard ratio 0.32, 95% CI 0.2–0.6, p = 0.001). Multifocality (hazard ratio 2.74, 95% CI 1.2–6.2, p = 0.02) was associated with worse BCSS, while progesterone receptor positivity (hazard ratio 0.27, 95% CI 0.1–0.7, p = 0.004), age at surgery (hazard ratio 0.94, 95% CI 0.9–1.0, p = 0.02 and BCS eligibility (hazard ratio 0.29, 95% CI 0.1–0.7, p = 0.009) were associated with improved BCSS. Type of breast surgery (BCS versus mastectomy) was not significantly associated with LRR, BCSS, or OS among this population. (Table 3). Given the multiple comparisons, Bonferroni correction was performed, and variables, including physical findings for OS, age, and multifocality for BCSS, and contralateral prophylactic mastectomy for LRR, were no longer significant after multiple comparison correction. Knowledge of genetic mutation status remained an independent beneficial factor for LRR, while BCS eligibility remained protective for OS and BCSS, likely reflecting differences in underlying prognostic factors.

TABLE 3.

Multivariable analysis

Locoregional Recurrence
	HR (95% CI)	p-value
Surgery		0.3
BCS	-
Mastectomy	2.42 (0.5–11.3)
Genetic status unknown at surgery	3.39 (1.4–8.8)	0.01
Contralateral prophylactic mastectomy	0.21 (0.05–0.95)**	0.04
Breast Cancer-Specific Survival
Surgery		0.4
BCS	-
Mastectomy	0.59 (0.2–1.8)
Age at surgery	0.94 (0.9–1.0)**	0.02
BCS eligible	0.29 (0.1–0.7)	0.009
Progesterone receptor positive	0.27 (0.1–0.7)	0.004
Multifocality	2.74 (1.2–6.2)**	0.02
Overall Survival
Surgery		0.4
BCS	-
Mastectomy	0.70 (0.3–1.6)
Presentation with physical findings	2.72 (1.2–6.2)**	0.02
BCS eligible	0.32 (0.2–0.6)	0.001
Receipt of neoadjuvant chemotherapy	2.61 (1.3–5.3)	0.008

^*None of the covariates were significant on multivariable analysis for contralateral breast cancers

^**These variables were not significant after multiple comparison correction

HR hazard ratio, CI confidence interval, BCS breast-conserving surgery

DISCUSSION

In our analysis of BRCA germline mutation carriers with an index breast cancer and close to 8 years of follow-up, there were no differences in rates of LRR, BCSS, and OS among women undergoing BCS or mastectomy. Patients aware of their BRCA genetic mutation status were more likely to opt for mastectomy than breast conservation at the time of breast cancer diagnosis. While there was a higher incidence of subsequent contralateral breast cancer in those who chose unilateral breast surgery, OS was unchanged.

In patients presenting with sporadic breast cancer, it is well established that BCS and mastectomy have equivalent OS and BCSS, and result in similar rates of local recurrence, supporting the routine use of BCS when feasible.^4,9,10 When counseling patients on a surgical approach, risk factors for subsequent breast cancer events are assessed. Our data and other studies demonstrate that women who are aware of their mutation status are significantly more likely to opt for mastectomy, both for therapeutic and prophylactic intent,¹¹ although it remains unclear if surgical management choice for an index cancer among BRCA mutation carriers improves outcomes.

When comparing rates of IBTR among women with and without a BRCA mutation following BCS, the data remain mixed. Studies by Robson et al. and Pierce et al. have demonstrated similar risk of IBTR in BRCA mutation carriers following BCS, while Haffty et al. reported significantly higher rates of IBTR following BCS among women with (49%) or without (21%) a BRCA mutation.^12,13

While the literature shows similar rates of local failure in the general breast cancer population following BCS and mastectomy with modern multimodality therapy, there are again mixed data on this topic among patients with a BRCA mutation. Pierce and colleagues reported on 655 women with BRCA mutations with long-term follow-up, and they found a significantly elevated rate of local failure following BCS compared with mastectomy, with a 15-year estimated risk of 23.5% versus 5.5%, respectively. Interestingly, the authors note that most events appeared to be second primary cancers rather than true local recurrences.¹⁴ In the current cohort of BRCA mutation carriers, we found no difference in rates of LRR among those treated with BCS or mastectomy. With longer follow-up, we may identify more second cancer events rather than true local recurrences; however, our more recent series may also reflect the ongoing trend in decreasing rates of local recurrence with ongoing improvements in adjuvant therapies.¹⁵

Patients with BRCA1 and/or BRCA2 genetic mutations are more likely to experience a contralateral breast cancer event.^2,7,12,14,16 Ye and colleagues reported that women with BRCA1 and/or BRCA2 genetic mutations were significantly more likely to experience a contralateral breast cancer following BCS, with 9.5% of carriers and only 0.7% of non-carriers having a CBC diagnosed in their median follow-up of 61 and 70 months, respectively (p < 0.001).¹⁷ When comparing those with BRCA1 versus BRCA2 mutations, Metcalfe et al. reported that the risk of contralateral breast cancer was 36.1% for women with a BRCA1 mutation and 28.5% for women with a BRCA2 mutation. Additionally, the risk of developing a contralateral breast cancer event was higher for women less than 50 years of age at the time of their index breast cancer diagnosis.¹⁶ Our data similarly show a relatively high rate of contralateral breast cancer among women undergoing unilateral breast surgery, with 10-year rates of contralateral breast cancer following unilateral surgery of 13.9%. Given this elevated predisposition to developing a contralateral breast cancer event, women who opt for unilateral breast surgery with a BRCA1 and/or BRCA2 mutation should be informed of their future risk and the importance of ongoing high-risk surveillance.¹⁸

Although the risk of developing breast cancer is higher in BRCA1 and/or BRCA2 mutation carriers than it is in the general population, it is unclear if undergoing mastectomy or contralateral risk-reducing surgery at the time of initial cancer diagnosis affects overall survival.^19,20 Despite a higher rate of contralateral breast cancer among women undergoing unilateral breast surgery, we found no difference in survival outcomes by surgery selection after 8 years of follow-up. Other studies have reported a potential survival benefit for women with a BRCA1 or BRCA2 pathogenic variant and a primary breast cancer opting for bilateral mastectomy. Metcalfe and colleagues report a 48% reduction in death from breast cancer at 20 years after contralateral mastectomy (p = 0.003).²¹. A recent systematic review by Davey et al., however, reported no difference in survival for patients undergoing BCS versus mastectomy for a primary breast cancer (hazard ratio 1.1, p = 0.66).¹⁹ Given conflict in the literature and our finding of no difference in OS between surgical management options for primary breast cancer, undertaking additional risk-reducing surgery with bilateral mastectomy is an individualized decision and should take into account risk assessment, including genetic results, age, tumor features, planned adjuvant therapy, and patient preference.

Our study has potential limitations, including those inherent to a retrospective review from a single institution which relies on appropriate follow-up and documentation of recurrent events. In addition, although our follow-up of 7.8 years is long relative to other series, and should capture most early local and distant recurrence events, additional in-breast second primary cancers or contralateral breast cancers may be detected with longer follow-up. This distinction remains an important data point, particularly when comparing outcomes among women with a pathogenic BRCA mutation. Furthermore, although 99 cancers were initially treated with BCS, 43 of those patients subsequently opted to undergo a contralateral prophylactic mastectomy more than 6 months after their initial surgery date, minimizing the overall BCS cohort over time.

Conclusion

Our analysis sought to compare a modern cohort of BRCA1 and/or BRCA2 mutation carriers receiving breast cancer therapy in the context of recent advances in screening, surgery, systemic therapy, and radiation technique. After nearly 8 years follow-up, we found no difference in rates of LRR, BCSS, and OS among BRCA mutation carriers following BCS or mastectomy for an index breast cancer, although longer follow-up is needed to assess long-term risk of second primary cancers. While patients who were aware of their BRCA mutation status were more likely to choose bilateral mastectomy at the time of their index cancer diagnosis, the only difference between surgical groups among our cohort of patients treated since 2006 was an increase in the rate of CBC for women undergoing unilateral surgery. Given that there was no detected difference in OS, women who desire to avoid mastectomy or those with competing morbidity may be counseled that breast conservation with continued high-risk surveillance is a reasonable treatment strategy for their breast cancer.

Synopsis:

Here we compare outcomes among BRCA mutation carriers undergoing BCS versus mastectomy. We conclude that BCS is an option for BRCA mutation carriers willing to continue high-risk screening.