Advances in the systemic treatment of triple-negative breast cancer.

Triple-negative breast cancer constitutes a heterogeneous group of malignancies that are often aggressive and associated with a poor prognosis. Molecular characterization, while not a standard of care, can further subtype triple-negative breast cancer and provide insight into prognostication and behaviour. Optimal chemotherapy regimens have yet to be established; however, there have been advances in the systemic treatment of triple-negative breast cancer in the neoadjuvant, adjuvant, and metastatic settings. In this review, we discuss evidence for the potential benefit of neoadjuvant platinum-based chemotherapy, adjuvant combination chemotherapy with weekly paclitaxel, and BRCA mutation-directed therapy in the metastatic setting. The role for adjuvant capecitabine in patients who do not achieve a pathologic complete response with neoadjuvant chemotherapy is reviewed. Future directions and data concerning novel targeted agents are reviewed, including the most recent data on parp [poly (adp-ribose) polymerase] inhibitors, antiandrogen agents, and immunotherapy.


INTRODUCTION
Breast cancer continues to be the most common solid tumour affecting women, and it is the second leading cause of cancer-related death in women 1 . In triple-negative breast cancer (tnbc), which accounts for approximately 10% -15% of diagnosed breast cancers 2 , expression of the estrogen and progesterone receptors is lacking, and the tumour is also negative for overexpression of her2 (human epidermal growth factor receptor 2) 3,4 .
Compared with the hormone receptor-positive breast cancers, tnbc has a worse prognosis, with an aggressive natural history 2 , and it is more commonly seen in younger and obese women, the average age of onset being 53 years. The prevalence of tnbc is higher in premenopausal African American women [5][6][7][8][9] . Importantly, tnbc has a strong correlation with BRCA1/2 mutation status, and up to 20% of tnbc patients will be carriers of a mutation 10 . Guidelines suggest screening for BRCA status in women with tnbc diagnosed at 60 years of age or younger (Table i).
Approximately 70% of tnbcs fall into the basallike subtype, and most basal-like cancers are triplenegative; however, those characteristics are not mutually exclusive 19 . In a recent genomic analysis of tnbc, four subtypes were described: luminal androgen receptor, mesenchymal, basal-like immunosuppressed, and basallike immunoactivated 20 . Of those subtypes, basal-like immunoactivated is associated with the best prognosis 20 , which is in keeping with prior research showing that prognosis is better for tnbc tumours with lymphocytic infiltration 21 . The BRCA-mutated cancers tend to be triplenegative and generally fall into the basal subtype 22 . Tumours that do not have germline mutations in BRCA1/2, but that display the characteristics of BRCA pathway deficiency are described as having "BRCAness" 23 . Those tumours are proposed to behave potentially more similarly to BRCA-mutated cancers in terms of natural history and response to systemic therapy. Molecular characterization of tnbc is an area of active research, but the application and relevance of that research to clinical practice has yet to be established.
At diagnosis, tnbc tumours are more likely to be T2 or T3, to be positive for lymphovascular invasion, and to have already metastasized to lymph nodes 5 . The pattern of spread is distinct from that for hormone receptor-positive tumours: tnbc has a greater propensity for brain and lung metastases, and a lower prevalence of bone metastases 8 . In a large observational prospective study of women with stages i-iii breast cancer, women with tnbc were found to have worse overall survival (os) compared with those having hormone receptor-positive, her2-negative tumours [hazard ratio (hr): 2.72; p < 0.0001] 8 . The difference was most pronounced in the first 2 years, the hr for os being 8.30 8

SYSTEMIC TREATMENT FOR TNBC Neoadjuvant Setting
Neoadjuvant chemotherapy is used in the treatment of localized early-stage breast cancer with a goal of breast-conserving surgery, or for patients for whom surgery is temporarily contraindicated 13,24 . The use of chemotherapy in the neoadjuvant setting allows for a direct assessment of the in vivo response by clinical examination or imaging evaluation. An Oncotype DX recurrence score that is associated with an estimated relapse risk of 15% or more at 10 years n HER2 positivity In patients who can tolerate it, use of an anthracycline-taxane regimen is considered the optimal strategy for adjuvant chemotherapy, particularly in patients deemed to be high risk.
Advanced or metastatic disease ESMO For triple-negative locally advanced BCa, anthracycline-taxane chemotherapy is recommended as initial treatment.

I, A
In triple-negative advanced BCa patients (regardless of BRCA status) previously treated with an anthracycline with or without a taxane in the neoadjuvant or adjuvant setting, carboplatin (compared with docetaxel) demonstrated comparable efficacy and a more favourable toxicity profile. It is therefore an important treatment option.

I, A
In patients with BRCA-associated triple-negative or endocrine-resistant metastatic BCa previously treated with an anthracycline with or without a taxane (in the adjuvant or metastatic setting, or both), a platinum regimen, if not previously administered, is the preferred option when no suitable clinical trial is available. Neoadjuvant chemotherapy results in higher rates of pathologic complete response (pcr) in tnbc than in hormone receptor-positive, her2-negative disease (28% -30% vs. 6.7%) 25 . The rate of pcr varies according to the subtype of tnbc, with the basal-like 1 subtype having the highest frequency of pcr (52%) and the basal-like 2 and luminal androgen receptor subtypes having the lowest frequency 26 . In a prospective database analysis, response to neoadjuvant therapy and long-term survival were compared in patients with tnbc and non-tnbc 27 . In that study, the rate of pcr was found to be higher in tnbc than in non-tnbc (22% vs. 11%, p = 0.34); however, tnbc was associated with a decreased 3-year progression-free survival (pfs) and a decreased 3-year os 27 . Patients who achieved a pcr showed the strongest association with positive long-term outcomes 25,27,28 .
The optimal chemotherapy regimen for the neoadjuvant treatment of tnbc has not been established. Platinum-based regimens have been suggested to possibly be more active in tnbc 29 . In the Cancer and Leukemia Group B 40603 (Alliance) study, the rate of pcr was compared in patients receiving carboplatin or bevacizumab (or both) in addition to weekly paclitaxel, followed by dose-dense doxorubicin and cyclophosphamide 30 . Rates of pcr were significantly improved with the addition of either carboplatin or bevacizumab in breast-confined disease (60% vs. 44%, p = 0.0018, and 59% vs. 48%, p = 0.0089, respectively). In locally advanced disease involving both breast and axilla, only carboplatin resulted in improved rates of pcr (54% vs. 41%, p = 0.0029).
In t he Gepa rSi x to gbg 66 study, pat ients w it h stages ii-iii tnbc or her2-positive breast cancer were treated with neoadjuvant weekly paclitaxel and non-pegylated liposomal doxorubicin, and either bevacizumab for tnbc or trastuzumab every 3 weeks for her2-positive breast cancer, with or without the addition of weekly carboplatin (area under the curve 1.5) 31 . The study found that pcr was observed more frequently in patients with tnbc who received additional carboplatin (53.2% vs. 36.9%, p = 0.005); however, that result came at the expense of greater toxicities.
As did the Cancer and Leukemia Group B 40603 study, the GeoarSixto gbg 66 study reported rates of pcr and did not assess disease-free survival (dfs) and os. However, pcr was associated with improved long-term outcomes 25,27 . Thus, although dfs and os were not studied, some experts believe that, given the increased rates of pcr, an os benefit can be predicted. However, controversy surrounds that assumption. As investigated in a meta-analysis by Cortazar et al. 28 , using pcr as a surrogate endpoint for event-free survival or os could not be validated.

Adjuvant Setting
As of February 2018, European Society for Medical Oncology guidelines do not recommend further adjuvant systemic treatment if residual disease is present after completion of neoadjuvant chemotherapy 13 . However, that principle has recently been challenged in the create-x trial 32 . In that study, patients with her2-negative disease who did not achieve a pcr with neoadjuvant chemotherapy were randomized to receive either standard of care, which included hormonal therapy or radiation therapy if indicated (or both), or the addition of oral capecitabine (1250 mg/m 2 twice daily for 14 of 21 days) for 6-8 cycles.
The study included patients with both hormone receptorpositive and -negative tumours.
The primary endpoint of the study was dfs, and the secondary endpoint was os. Collectively, statistically significant improvements in dfs and os were observed (74.1% vs. 67.6%, p = 0.01, and 89.2% vs. 83.5%, p = 0.01, respectively). A hr of 0.59 (95% ci: 0.39 to 0.90; p = 0.01) was reported for os. On subgroup analysis, the greatest benefit was observed in patients with tnbc, with the dfs rate being 69.8% in the capecitabine group compared with 56.1% in the standard-treatment group. Similarly, os was greater for patients with tnbc in the capecitabine arm (78.8% vs. 70.3%; hr: 0.52; 95% ci: 0.30 to 0.90). Adverse events were frequent in patients taking capecitabine, with 73.4% reporting hand-foot syndrome. Dose reductions were required in 23.9% and 36.7% of the patients assigned to 6 and 8 cycles of capecitabine respectively. In contrast to the European Society for Medical Oncology guidelines, the U.S. National Comprehensive Cancer Network guidelines were updated in February 2018 and incorporate the consideration of using capecitabine in this setting 14 .
A number of studies have examined the potential benefit of adjuvant treatment after neoadjuvant chemotherapy. They include a phase iii study examining the use of avelumab, a monoclonal antibody inhibitor of PD-L1, in the adjuvant or post-neoadjuvant setting in high-risk patients (see NCT02926196 at http://ClinicalTrials.gov) and another phase iii trial assessing pembrolizumab in patients with tnbc who have residual disease after neoadjuvant chemotherapy (see NCT02954874). Table ii presents studies for tnbc patients that are currently recruiting in Canada.
For patients who do not receive neoadjuvant chemotherapy (with the possible exception of those having rare histologic subtypes), the European Society for Medical Oncology guidelines suggest treatment with adjuvant chemotherapy ( Table i) 13 . Some controversy surrounds the choice of systemic chemotherapy for small tumours (≤0.5 cm) that are node-negative, and that decision must therefore be individualized [13][14][15][16] . The optimal adjuvant regimen for tnbc has not been established, but current guidelines support the use of regimens that contain an anthracycline and a taxane, if feasible (Table i) 13,16 .
The geicam 9906 trial compared adjuvant fluorouracilepirubicin-cyclophosphamide (fec) with fec-p (fec followed by weekly paclitaxel) in lymph node-positive breast cancer 33 . That study found a 23% reduction in the risk of relapse and a 22% reduction in the risk of death with the addition of paclitaxel. On subgroup analysis, patients with tnbc were found to experience improved dfs when treated with fec plus weekly paclitaxel compared with fec alone (76% vs. 62%, p = 0.0254) 34 .
The role of weekly paclitaxel was also studied in the E1199 phase iii trial 35 . In that study, women with stages ii-iii breast cancer were treated with 4 cycles of doxorubicincyclophosphamide followed by paclitaxel or docetaxel every 3 weeks for 4 doses or weekly for 12 doses in a 2×2 design 35 . At the 10-year follow-up, significant improvement in dfs and os was observed for the tnbc subgroup treated with weekly paclitaxel compared with paclitaxel every 3 weeks (hr: 0.69; p = 0.01) or with docetaxel in either  schedule (hr: 0.69; p = 0.019). Those findings suggest that a benefit might accrue to the addition of weekly paclitaxel to adjuvant chemotherapy in tnbc; however, that regimen was not the primary objective of the study, and thus it is difficult to base recommendations on the subgroup analysis alone. The addition of bevacizumab to chemotherapy was studied in the adjuvant setting in the beatrice study 36 . No invasive dfs or os benefit was demonstrated in that setting.
Although studies supporting the role of neoadjuvant and palliative platinum-based chemotherapy in tnbc have been published, no data in the adjuvant setting are currently available. Clinical trials investigating the role of platinum-based adjuvant chemotherapy are ongoing (for example, see NCT02488967 at http://ClinicalTrials. gov; Table ii).

Metastatic Setting
Triple-negative breast cancer is associated with a high risk of distant recurrence, predominantly in the first 2 years after diagnosis 5,8 . When metastases occur, biopsy of the site of distant disease should be attempted to assess for discordance in hormone receptor and her2 status 37 . A retrospective analysis found that 8% of tumours that were initially estrogen receptor-negative had converted to estrogen receptor-positivity when the metastatic tumour deposit was assessed for hormone receptor status 38 . No statistical discordance in her2 receptor status was observed.
The choice of initial systemic chemotherapy should be individualized based on a number of factors, including tumour burden, rate of disease progression, performance status, previous chemotherapy exposure, and patient preferences 17,18 . With respect to agent selection, Table i describes several guideline recommendations. Although combination chemotherapy is generally avoided in the palliative setting, tnbc often results in visceral involvement and a more aggressive course, making combination chemotherapy a more frequent choice in this population 39 .
Platinum-based chemotherapy has been suggested to potentially be more effective than non-platinum-based chemotherapy in metastatic tnbc. In a retrospective cohort study, longer pfs was observed in patients receiving platinum-based chemotherapy compared with nonplatinum-based therapy in the first line (7.8 months vs. 4.9 months, p < 0.001) 40 . However, no difference in os was observed. No improvement in pfs or os was shown in the preliminary results of the prospective tnt study, which compared carboplatin with docetaxel in metastatic or recurrent locally advanced tnbc 41 . The tnt study randomized 376 patients with metastatic tnbc or with known BRCA1/2 mutation to either carboplatin (area under the curve 6 every 3 weeks) or docetaxel (100 mg/m 2 every 3 weeks) for 6-8 cycles or until progression. The initial results, presented at the 2014 San Antonio Breast Cancer Symposium, showed no statistical difference in pfs or os for the tnbc group (3.1 months vs. 4.5 months and 12.4 months vs. 12.3 months respectively). However, on subgroup analysis of BRCA1/2 carriers, pfs was improved in those receiving carboplatin (6.8 months vs. 3.1 months). The objective response rate was also significantly higher (68.0% vs. 33.3%, p = 0.03). Interestingly, on subgroup analysis of basal-like malignancies by immunohistochemistry, no statistical difference in the overall response rate was evident between the two treatment arms. That finding suggests that "BRCAness" might not be a reliable means of predicting the clinical response of tnbcs to certain chemotherapy agents.

BRCA Carriers and TNBC
The treatment of BRCA mutation-associated breast cancer and the use of directed agents for that patient subset is an active area of research. Because the BRCA1 and BRCA2 genes code for proteins involved in double-stranded dna break repair, it is hypothesized that BRCA mutationassociated cancers might be more sensitive to chemotherapy agents that cause dna damage, such as the platinums. Sensitivity of that kind has been shown in vitro 42 ; however, its translation into clinical application has yet to be established. A small phase ii study of cisplatin chemotherapy in BRCA1 mutation carriers, not selected for hormone receptor status, demonstrated high efficacy for cisplatin, with an overall response rate of 80% and 9 of 20 patients achieving a complete response 43 . However, the study lacked a control group. Further research into the role of cisplatin in the treatment of metastatic tnbc and BRCA mutationassociated breast cancer is ongoing (NCT02595905 at http://ClinicalTrials.gov).
The impaired dna repair pathway in BRCA1/2 carriers can also be targeted with parp [poly (adp-ribose) polymerase] inhibitors, which interfere with the repair of single-stranded dna breaks and, when combined with an already weakened repair process, can result in synthetic lethality 44 . That activity has been confirmed by in vitro studies, showing that tumours in carriers are in fact sensitive to parp inhibitors 45,46 . Recently, the parp inhibitor olaparib was studied in the setting of metastatic breast cancer in the olympiad trial 47 . That study included 302 patients with metastatic breast cancer who had known BRCA mutations, who were negative for her2 overexpression, and who had received up to 2 prior lines of chemotherapy. The patients were randomized to single-agent chemotherapy (capecitabine, eribulin, or vinorelbine every 3 weeks) or to olaparib (300 mg twice daily). The response rate was 59.9% in the olaparib group compared with 28.8% in the chemotherapy group. Furthermore, pfs was improved in the olaparib group (7.0 months vs. 4.2 months; hr: 0.58; 95% ci: 0.43 to 0.80; p < 0.001). Olaparib was approved in January 2018 by the U.S. Food and Drug Administration for use in BRCA-mutated metastatic breast cancer. In Canada, olaparib is currently approved for use only in BRCA-mutated ovarian, fallopian tube, or primary peritoneal cancers. A phase iii Canadian Cancer Trials Group study, currently open to accrual, is examining the role of olaparib in the adjuvant setting for carriers of BRCA1/2 mutations (see NCT02032823 at http://ClinicalTrials.gov). Other parp inhibitors are also actively being investigated in the metastatic setting, including veliparib and niraparib (for example, NCT02595905, NCT01905592) 48 . Whether tumours that are BRCA-proficient or that have BRCA pathway impairment (BRCAness) will respond to these targeted therapies is currently unknown.
The angiogenesis inhibitor bevacizumab has been examined in combination with chemotherapy agents in the treatment of metastatic breast cancer 49,50 . Although the relevant studies demonstrated benefit in pfs, no improvement in os was observed. Miles et al. 51 reported a pooled subgroup analysis of bevacizumab use in poor-prognosis groups. In the tnbc subgroup, median pfs was significantly improved with the use of bevacizumab (hr: 0.63; 95% ci: 0.53 to 0.76) 51 ; however, that improvement did not translate into an os benefit. Thus, no role for bevacizumab is accepted at this time.

FUTURE DIRECTIONS: IMMUNOTHERAPY AND TARGETED THERAPY
The role of immunotherapy in the treatment of tnbc is currently under investigation in several trials recruiting in Canada (Table ii). In tnbc, strong lymphocytic infiltration or immune response has been associated with improved prognosis 52 . That observation suggests that harnessing the immune system against this disease might be beneficial. Pembrolizumab is a PD-1 inhibitor that has been shown to be effective in the treatment of several other cancers, including lung, melanoma, and bladder malignancies [53][54][55] . In a phase ib study of pembrolizumab in patients with advanced tnbc, only 18.5% of patients experienced a complete or partial response 56 , with the duration of response varying from 15 to more than 47.3 weeks. A phase iii trial of pembrolizumab in the treatment of metastatic or locally recurrent inoperable tnbc is ongoing (see NCT02819518 at http://ClinicalTrials.gov) 57 . The anti-PD-L1 antibody atezolizumab is also actively being investigated in the neoadjuvant and metastatic settings (see NCT03197935 and NCT03125902). Immunotherapy has shown great potential in a number of other disease sites, and thus the results of the foregoing trials are highly anticipated.
In a subgroup of tnbc, expression of the androgen receptor (ar) is increased 20 . The clinical relevance of ar status has yet to be established in breast cancer; however, several recent phase ii studies have examined the use of anti-androgen agents in this setting. In one study, tumours from 242 patients with breast cancer negative for the estrogen and progesterone receptors were tested for ar expression 58 . Of those tumours, only 12% were found to be ar-positive. Patients whose tumours tested positive received bicalutamide 150 mg daily. The clinical benefit rate, defined as complete response, partial response, or stable disease for more than 6 months, was 19%, and the median pfs was 12 weeks (95% ci: 11 weeks to 22 weeks) 58 . A phase iii evaluation of bicalutamide is still pending.
Enzalutamide is an ar signalling inhibitor that is used in the treatment of metastatic castration-resistant prostate cancer. The effect of enzalutamide has also been studied in ar-positive tnbc in a phase ii trial 59 . In that trial, androgen positivity was defined as any level exceeding 0% by immunohistochemistry, and patients were further assessed for an androgen-driven gene signature by gene profiling. The clinical benefit rate was greater in patients with a positive gene signature (39% vs. 11%). Additionally, median pfs was 32 weeks compared with 9 weeks for patients testing negative. Research into the role of enzalutamide in the treatment of ar-positive tnbc is being conducted in the United States (see NCT02750358 at http://ClinicalTrials.gov).

SUMMARY
Triple-negative breast cancer constitutes a heterogeneous group of malignancies that differ in natural history and response to treatment. The mainstay of treatment continues to be chemotherapy; however, optimal chemotherapy regimens for tnbc have yet to be established.
In this article, we have reviewed the current evidence for systemic treatment for tnbc in the neoadjuvant, adjuvant, and metastatic settings. In the neoadjuvant setting, the use of platinum agents has been associated with improved rates of pcr, but os was not reported in the associated studies 30,31 . In the recently published create-x trial, an os benefit was shown for adjuvant capecitabine in patients who do not achieve a pcr with neoadjuvant chemotherapy 32 . However, that approach has not been widely adopted in clinical practice, perhaps because of the associated toxicity. For adjuvant treatment of tnbc, two trials that added weekly paclitaxel to combination chemotherapy showed improvements in dfs [33][34][35] . In metastatic tnbc, preliminary results show that the response rate might be higher with platinum than with docetaxel in BRCA mutation-associated malignancies 40 ; however, research to guide the optimal choice of systemic treatment for metastatic disease is limited. Olaparib, a parp inhibitor, has recently been approved by the U.S. Food and Drug Administration for use in germline BRCA-mutated metastatic breast cancer, based on the olympiad trial 47 .
As molecular research advances an understanding of the driver mutations in this disease, more targeted treatments could become available. A number of investigational therapies hold promise, including parp inhibitors, ar pathway inhibitors, and immunotherapy. Given those new developments, the hope is that more effective treatments and better outcomes will be achieved for patients with tnbc.

CONFLICT OF INTEREST DISCLOSURES
We have read and understood Current Oncology's policy on disclosing conflicts of interest, and we declare the following interests: RL has received education funding from Amgen and Eisai and fees as an advisory board member for AstraZeneca, Novartis, and Pfizer. EP has received travel funding from Roche and fees as an advisory board member for Amgen, Genomic Health, Novartis, and Pfizer. MS has received conference funding from Amgen and fees as an advisory board member for Pfizer and Shire. JM has received conference travel funding from AstraZeneca and Roche and fees as an advisory board member from Novartis and Pfizer.

AUTHOR AFFILIATIONS
*Memorial University of Newfoundland, Saint John's, NL.