Clinical behavior and outcomes of breast cancer in young women with germline BRCA pathogenic variants

Young breast cancer (BC) patients carrying a germline BRCA pathogenic variant (mBRCA) have similar outcomes as non-carriers. However, the impact of the type of gene (BRCA1 vs. BRCA2) and hormone receptor status (positive [HR+] vs. negative [HR−]) on clinical behavior and outcomes of mBRCA BC remains largely unknown. This is an international, multicenter, hospital-based, retrospective cohort study that included mBRCA patients diagnosed, between January 2000 and December 2012, with stage I–III invasive early BC at age ≤40 years. From 30 centers worldwide, 1236 young mBRCA BC patients were included. Among 808 and 428 patients with mBRCA1 or mBRCA2, 191 (23.6%) and 356 (83.2%) had HR+tumors, respectively (P < 0.001). Median follow-up was 7.9 years. Second primary BC (P = 0.009) and non-BC malignancies (P = 0.02) were more frequent among mBRCA1 patients while distant recurrences were less frequent (P = 0.02). Irrespective of hormone receptor status, mBRCA1 patients had worse disease-free survival (DFS; adjusted HR = 0.76, 95% CI = 0.60–0.96), with no difference in distant recurrence-free interval (DRFI) and overall survival (OS). Patients with HR+ disease had more frequent distant recurrences (P < 0.001) and less frequent second primary malignancies (BC: P = 0.005; non-BC: P = 0.18). No differences in DFS and OS were observed according to hormone receptor status, with a tendency for worse DRFI (adjusted HR = 1.39, 95% CI = 0.94–2.05) in patients with HR+ BC. Type of mBRCA gene and hormone receptor status strongly impact BC clinical behavior and outcomes in mBRCA young patients. These results provide important information for patients’ counseling on treatment, prevention, and surveillance strategies.


INTRODUCTION
In women aged ≤40 years, breast cancer is the most common malignancy and the first cause of cancer-related mortality 1 . Despite the higher risk of developing triple-negative and HER2-positive Other age-related issues should be considered in the management of breast cancer in young women 4 . Among them, genetic counseling and testing is key. Approximately 12% of cases arising in women aged ≤40 years are related to germline pathogenic variants in BRCA1 or BRCA2 5,6 . Indeed, young age at diagnosis is a criterion to refer patients to genetic counseling irrespective of family history or other tumor biological features 4 . Identification of a germline pathogenic variant in the BRCA genes plays a crucial role in cancer prevention and treatment 7,8 . BRCA-related breast cancers have distinct biological features, including a tendency for hormone receptor-negativity in BRCA1 carriers and hormone receptor positivity in BRCA2 carriers [9][10][11][12] .
Several studies have investigated the prognostic role of carrying germline BRCA pathogenic variants 13 . Compared to women with sporadic breast cancer, current evidence does not support different outcomes in those with BRCA pathogenic variants 13 . Similarly, no difference in survival outcomes between young breast cancer patients with or without germline BRCA pathogenic variants have been shown, except a trend for a survival advantage in BRCA-mutated patients with triple-negative breast cancer compared with non-carriers 6 . Nevertheless, besides the known differences in clinicopathological features in breast cancer cases associated with germline BRCA1 or BRCA2 pathogenic variants [9][10][11][12] , there is a lack of evidence so far on whether the type of mutated gene may also lead to potential differences in breast cancer clinical behavior and outcomes. In addition, the prognostic value of hormone receptor status in BRCA-related breast cancers remains largely unknown. We addressed these important issues in a large series of young breast cancer patients harboring germline BRCA pathogenic variants.

Study cohort
Out of 1424 patients registered in the study from 30 centers worldwide, 1236 young BRCA-mutated breast cancer patients were eligible for inclusion in the present analysis ( Supplementary  Fig. 1).
Comparison between patients with germline BRCA1 and BRCA2 pathogenic variants Compared to patients in the BRCA2 cohort, those in the BRCA1 cohort were younger, more likely to have Israeli origin, had more grade 3 tumors, less lobular histology, nodal involvement, and HER2-positive disease (Table 1). Women in the BRCA1 cohort received chemotherapy more frequently and, among those with hormone receptor-positive disease, fewer patients received adjuvant endocrine therapy (Table 1). Although patients in the BRCA1 cohort underwent breast-conserving surgery more often, no difference was observed in rates of risk-reducing mastectomy or salpingo-oophorectomy between the BRCA1 and BRCA2 cohorts, respectively (Supplementary Table 1).
Supplementary Tables 2 and 3 report the comparison in clinicopathological characteristics and risk-reducing surgery between the BRCA1 and BRCA2 cohorts according to hormone receptor status.
Median follow-up was 7.9 years (interquartile range 5.6-10.6 years), with no difference between the BRCA1 and BRCA2 cohorts (P = 0.95). In terms of the first DFS event, second primary malignancies (breast cancer: 17.0% vs. 12.2%, P = 0.009; nonbreast cancer: 4.3% vs. 1.9%, P = 0.02) were more frequent in the BRCA1 cohort while distant recurrences were less frequent (10.4% vs. 15.4%, P = 0.02) as compared to the BRCA2 cohort (Table 2). When the pattern of the first DFS event was assessed between the BRCA1 and BRCA2 cohorts according to hormone receptor status, the only difference was a higher rate of second non-breast cancer primary malignancies (5.2% vs. 1.4%, P = 0.005) in the BRCA1 cohort with hormone receptor-positive disease, being mostly ovarian cancer (Supplementary Table 4).
Considering DFS events, a similar pattern of annual HRs was observed between patients in the BRCA1 and BRCA2 cohorts irrespective of hormone receptor status, with a higher risk for those with BRCA1 pathogenic variants (Figs. 1a, 2a, and 3a).
Eight-year DFS was 62.8% and 65.9% in the BRCA1 and BRCA2 cohorts, respectively (adjusted HR = 0.76, 95% CI 0.60-0.96; Fig. 1b). A tendency towards worse DFS in the BRCA1 cohort was observed for both patients with hormone receptor-positive (adjusted HR = 0.77, 95% CI 0.58-1.03; Fig. 2b) and negative (adjusted HR = 0.73, 95% CI 0.48-1.13; Fig. 3b) breast cancer (P interaction = 0.85; Supplementary Comparison between patients with hormone receptorpositive and negative disease Compared to patients with hormone receptor-negative disease, those with hormone receptor-positive breast cancer were older, had less grade 3 tumors, more often lobular histology, nodal involvement, and HER2-positive disease (Supplementary Table 8). Women with hormone receptor-positive breast cancer were less likely to receive chemotherapy and underwent mastectomy more often than those with hormone receptor-negative disease (Supplementary Table 8). No differences were observed in rates of risk-reducing mastectomy and salpingo-oophorectomy between patients with hormone receptor-positive and negative breast cancer, respectively (Supplementary Table 9).
In terms of first DFS event, patients with hormone receptorpositive breast cancer had a higher incidence of distant (±locoregional) recurrences (16.1% vs. 9.0%, P < 0.001) and lower incidence of second primary malignancies (breast cancer: 12.1% vs. 17.9%, P = 0.005; non-breast cancer: 2.8% vs. 4.0%, P = 0.18) compared to women with hormone receptor-negative disease ( Table 3). A similar pattern of annual HRs was observed between patients with hormone receptor-positive and negative disease until 5 years, beyond which a rapidly increasing trend for those with hormone receptor-positive breast cancer was observed (Fig.  4a).
Considering that the cut-off used for defining hormone receptor positivity was not homogenous in all centers, the analyses comparing between patients with hormone receptor-positive and negative disease were then repeated by including only patients for whom the 1% cut-off for estrogen and/or progesterone receptor expression in their tumor was used to define hormone receptor status. Results were consistent with those reported in the main analyses (Supplementary Tables 11-14 and Supplementary Figs. [2][3][4][5].

DISCUSSION
This large unique dataset allowed an in-depth investigation of the impact of the type of mutated gene (BRCA1 vs BRCA2) and hormone receptor status on clinical behavior and outcomes of BRCA-mutated breast cancer in young women. We observed that patients with germline BRCA1 pathogenic variants had worse DFS than those with BRCA2 pathogenic variants, mainly due to higher rates of second primary malignancies (predominantly breast and ovarian cancers) irrespective of hormone receptor status, with no difference in DRFI and OS. Hormone receptor positivity was not associated with favorable prognosis in BRCA-mutated breast cancer, whatever the type of mutated gene, but rather it displayed a tendency for worse DRFI and no difference in DFS and OS compared to hormone receptor-negative disease. These data are highly relevant for counseling young BRCA-mutated breast cancer patients.
In terms of clinicopathological features, as shown in prior studies [9][10][11][12] , the majority of breast cancer cases associated with germline BRCA1 pathogenic variants were hormone receptornegative tumors (76%) while those arising in BRCA2-mutated patients were hormone receptor-positive (83%). Overall, the majority of tumors were high-grade (60-80%) and HER2-negative (90-95%), both features being particularly relevant in the case of BRCA1 carriers [9][10][11][12]14 . In terms of clinicopathological characteristics in the BRCA1 and BRCA2 cohorts according to hormone receptor status, the features that remained differently distributed were country of origin and tumor grade in patients with hormone receptor-positive disease (more likely Israeli origin and high-grade tumors in the BRCA1 cohort) as well as age and HER2 status in those with hormone  receptor-negative tumors (younger age and more likely HER2negative status in the BRCA1 cohort).
Harboring a germline pathogenic variant in BRCA1 or BRCA2 was not previously shown to have an apparent negative prognostic impact in breast cancer 13 , neither in young patients 6 . Compared with non-carriers, a tendency for better survival outcomes in BRCA carriers with triple-negative breast cancer was shown 6 . In addition, a reverse association of the prognostic value of hormone receptor status was observed between breast cancer patients with or without germline BRCA pathogenic variants 6,15 , particularly in the case of BRCA2 pathogenic variants [16][17][18] . Nevertheless, within the cohort of young BRCA-mutated breast cancer patients, the prognostic implications of carrying a germline BRCA1 or BRCA2 pathogenic variant as well as potential differences according to hormone receptor status remained largely undefined. Two main findings with potential clinical implications were found in our study.
Firstly, irrespective of hormone receptor status, the type of BRCA gene does not appear to have prognostic value, with no observed difference in DRFI and OS between BRCA1-and BRCA2-mutated  patients. This is reassuring evidence for counseling young BRCAmutated breast cancer patients. Nevertheless, BRCA1 carriers showed worse DFS, mostly due to higher rates of second primary malignancies (predominantly breast and ovarian cancers) as compared to BRCA2-mutated patients. This result reflects the different and age-dependent risk of cancer development associated with the two genes: the breast cancer peak of incidence occurs earlier and ovarian cancer risk is higher in BRCA1 carriers as compared to those with BRCA2 pathogenic variants 19,20 . Young age at cancer diagnosis further increases the risk of second primary breast and ovarian malignancies 19,[21][22][23] . Therefore, these findings highlight the key role of BRCA genetic testing in this patient population and calls for awareness of the high risk of secondary malignancies and the option of considering riskreducing surgeries when a pathogenic variant is identified, particularly in BRCA1 carriers 24,25 . In addition, these results pinpoint the need to consider age-specific risk estimates when offering these procedures during patients' counseling 23 . Secondly, irrespective of the type of BRCA-mutated gene, hormone receptor positivity has no favorable prognostic value in this setting. While no difference in DFS and OS was observed, patients with hormone receptor-positive breast cancer showed a higher risk of distant relapses and a trend for worse DRFI as compared to those with hormone receptor-negative disease. There are several potential explanations for this lack of prognostic advantage in young BRCA-mutated patients with hormone receptor-positive breast cancer. Estrogen signaling was shown to promote tumor initiation and progression in BRCA-deficient cells 26,27 . Biological differences were reported between hormone receptor-positive breast cancers arising in patients with or without germline BRCA pathogenic variants 28,29 . When tested with OncotypeDx, a larger proportion of BRCA-mutated patients were found to have high recurrence scores, and low recurrence score was not necessarily associated with low risk of recurrence [30][31][32] . Taking into account that our study included only patients ≤40 years at diagnosis, these considerations may be further amplified due to the additional effect of young age on the biology of hormone-receptor-positive breast cancers 2,3 . Finally, it should be highlighted that BRCA-related tumors are known to have high chemosensitivity 11,17,33 , particularly in the case of hormone receptor-negative disease 34,35 . In our study, more than 90% of the patients received chemotherapy. However,~8% of patients with hormone receptor-positive disease did not receive adjuvant endocrine therapy, with significantly higher numbers of nonrecipients among BRCA1 carriers. Although the reasons for lack of endocrine therapy administration were not collected, special attention should be paid in this regard when counseling BRCAmutated patients with hormone receptor-positive disease. Our findings provide indirect evidence to potentially consider BRCAmutated patients with hormone receptor-positive disease as a high-risk population 36 . Therefore, ovarian function suppression 37 and extended adjuvant endocrine therapy 38 may be considered relevant options in this setting. In light of the increased ovarian cancer risk, risk-reducing salpingo-oophorectomy might be considered as the primary strategy for ovarian function The findings of this study should be considered in the context of its limitations. This is a retrospective cohort study conducted over a relatively long period of time. Assessment of BRCA and hormone receptor status as well as patient management were conducted according to diagnostic and treatment procedures available at that time in the respective country and center. Median follow-up was shorter than 10 years. Nevertheless, several unique features of this study should also be highlighted. This is a multicenter study, not restricted to a single continent, that included a large sample size despite focusing on a special and rare patient population (i.e., young breast cancer patients with germline BRCA pathogenic variants). The numbers of included patients and registered events made possible the acquisition of  reliable results on these important unmet and clinically highly relevant issues.
In conclusion, type of BRCA-mutated gene and hormone receptor status strongly impact the clinical behavior and outcomes of breast cancer in young patients with germline BRCA pathogenic variants. Young patients with germline BRCA1 pathogenic variants had worse DFS than those with germline BRCA2 pathogenic variants mostly due to higher rates of second primary malignancies irrespective of hormone receptor status. On the other hand, unlike breast cancers arising in non-BRCA carriers, hormone receptor status had no prognostic value in young BRCAmutated patients, and even a tendency of worse DRFI in women with hormone receptor-positive disease was observed. These results provide important information for counseling young BRCAmutated breast cancer patients regarding treatment, prevention and surveillance strategies.

Study design and participants
Details of this study were previously reported 39 . Briefly, this was an international, multicenter, hospital-based, retrospective cohort study that included women diagnosed at age ≤40 years with invasive early breast cancer (stage I-III) between January 2000 and December 2012. All included patients had a known germline BRCA1 or BRCA2 pathogenic variant.
Healthy carriers as well as women with BRCA variants of uncertain significance, other malignancies (including ovarian cancer) without a prior diagnosis of invasive breast cancer, in situ or stage IV de novo breast cancer, or lack of information on follow-up were not eligible for inclusion. For the purpose of the present analysis, patients harboring pathogenic variants in both BRCA1 and BRCA2 as well as those with unknown hormone receptor status were also excluded.
Datasets from countries with more than one participating center were crosschecked to exclude potential duplicated cases.

Procedures
Data on tumor and patient characteristics, treatment, BRCA pathogenic variants, and survival outcomes were collected for all eligible patients.
The type of mutated gene was the criteria used to distinguish between two cohorts of patients: women with BRCA1 (BRCA1 cohort) and those with BRCA2 (BRCA2 cohort) pathogenic variants.
BRCA pathogenic variants and hormone receptor status were assessed locally at each participating center. Hormone receptor positivity was defined by the expression of estrogen and/or progesterone receptors in ≥1% of invasive tumor cells (≥10% in French participating centers) assessed by immunostaining.
The Institut Jules Bordet (Brussels, Belgium) coordinated the study and acted as the central ethics committee. Ethics approval by the Institutional Review Boards of participating centers and patients' written informed consent were obtained before inclusion whenever requested by local regulations.
The STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) statement was followed for study reporting 40 .

Outcomes
The current analysis aimed to investigate the impact of the type of mutated gene (BRCA1 vs. BRCA2) and hormone receptor status on clinical behavior and outcomes of young breast cancer patients with germline BRCA pathogenic variants.
Clinicopathological characteristics, pattern, and risk over time of diseasefree survival (DFS) events, as well as prognosis (in terms of DFS, distant recurrence-free interval [DRFI], and overall survival [OS]) were compared between the BRCA1 and BRCA2 cohorts. The same analyses comparing the BRCA1 and BRCA2 cohorts were then performed separately in patients with hormone receptor-positive and negative disease.
To specifically assess the effect of hormone receptor status, clinicopathological characteristics, pattern, and risk over time of DFS events, as well as prognosis (in terms of DFS, DRFI, and OS) were compared between patients with hormone receptor-positive and negative disease irrespective of the type of mutated gene.

Statistical analysis
Descriptive analyses were used to assess clinicopathological characteristics as well as the pattern of DFS events.
To assess the risk of developing DFS events over time, the Epanechnikov Kernel-Smoothed annual hazards of recurrence were calculated.
DFS was defined as the time from diagnosis until the first appearance of one of the following invasive events: loco-regional recurrence, distant metastases, new contralateral or ipsilateral breast cancer, second primary malignancy, or death from any cause. DRFI was calculated as the time from diagnosis until the first occurrence of invasive breast cancer recurrence at a distant site. OS was defined as the time from diagnosis until death from any cause.
Observation times of patients that did not experience an event were censored on the date of their last contact. Kaplan-Meier plots were used to present results with a follow-up time of up to 10 years. Cox proportional hazard model was applied to estimate the hazard ratios (HRs) over the whole follow-up period, adjusting for the concomitant effect of selected confounders. Multivariate models for all survival analyses included nodal status, grade, HER2, type of breast surgery, chemotherapy use, age, year of diagnosis, and country.
Homogeneity tests on the HRs were performed in all survival analyses using the likelihood ratio test to assess whether there was evidence of an interaction between the type of gene and hormone receptor status.
All statistical analyses were two-sided; P values < 0.05 were considered as statistically significant. Statistical analyses were performed by MC and MB using Stata 13.1 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP).

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

DATA AVAILABILITY
The data generated and analysed during this study are described in the following data record: https://doi.org/10.6084/m9.figshare.13507422 41 . The data are stored in the following four Excel spreadsheets: 'Patient survival data.xlsx', 'Patient baseline, tumor and treatment data.xlsx', 'Patient risk-reducing surgery data.xlsx', 'Patient eligibility criteria and survival data.xlsx'. These data are not publicly available for the following reason: data contain information that could compromise research participant privacy. However, the data can be made available upon reasonable request to the corresponding author. A list of which data file underlies which figure, table, and supplementary table in the related manuscript is provided in the file 'Lambertini et al. underlying data lookup.csv', included as part of the metadata record 41 . The dataset analysed during this study is described with more details in the following manuscript: https://doi.org/10.1200/JCO.19.02399 39 .