Triple negative breast cancer (TNBC) is a clinical group of breast invasive tumors that account for approximately 15–20% of all breast cancers, both worldwide and in Brazil (1, 2). This group of tumors is clinically defined as both the absence of estrogen and progesterone receptors plus the absence of HER2 overexpression/amplification. Clinically, TNBC tends to present as fast growing tumors, often diagnosed as locally advanced.

Taken as a group, locally advanced breast cancers (LABC), generally defined as stage III tumors, account worldwide for about 10–30% of all breast cancers (3, 4). In some settings, as in Brazil, the proportion of patients with LABC is in the upper estimate, with approximately 25% of breast cancers presenting with stage III disease (and up to 47% with stage II disease) (2). The conjunction of TNBC biology and locally advanced presentation poses a clinical challenge: TNBC presents a high response to chemotherapy, but the likelihood of relapse remains high (5, 6). Therefore, systemic therapy before surgery (primary systemic therapy) is the most appropriate approach for patients with LABC, especially of TNBC biology, given the improved rate of inoperable tumors that become subsequently operable, the proportion of patients not candidates to breast-conserving surgery that are afterward amenable to receive such procedure, but also given the prognostic implications of the complete response to neoadjuvant chemotherapy, referred as pathological complete response (pCR). In fact, pCR emerged as a strong marker of prognosis, as patients attaining a complete response to primary systemic therapy have improved outcomes both in terms of recurrence and survival (7). This led both the Food and Drug Administration and the European Medicines Agency to recognize the neoadjuvant setting and the outcome pCR as an arena for the evaluation of the performance of different treatment regimens. Thus, in light of the high risk of relapse of TNBC and locally advanced staging, trials aiming at improving the efficacy of neoadjuvant treatments are a clear unmet need and an evolving field in current clinical practice.

In this setting, DNA cross-linking platinum salts [side with poly (ADP-ribose) polymerase—PARP—inhibitors in the metastatic setting] are being used as strategies to improve outcomes in TNBC (8, 9). These tumors present a high mutation load and features of genomic instability that are frequently associated with homologous recombination deficiencies (the most frequent cause being BRCA1 deficiency) (10, 11). The use of platinum salts aims at exploiting homologous recombination deficiency by further challenging this set of pathways in order to achieve synthetic lethality (12). In fact, several studies have now tested platinum salts (both carboplatin and cisplatin) in the neoadjuvant and palliative setting in TNBC (8, 13–16). Data on the use of cisplatin is, however, scarce. In a trial that enrolled 28 patients with TNBC treated with neoadjuvant single agent cisplatin for four cycles, 6 (22%) of 28 achieved a pCR, with 18 (64%) attaining a clinical complete or partial response. Cisplatin in combination with doxorubicin and cyclophosphamide (CAP) was already shown to be an effective and safe regimen to treat ovarian cancer (17–19). Its efficacy in patients with breast cancer is, however, unknown, but CAP regimen is an appealing approach given its standard anthracycline–cyclophosphamide backbone. Moreover, despite the risks for increased toxicity, the combined use of the three agents might further improve response and the proportion of patients reaching a pCR.

In this prospective, single arm, phase II study, we sought to evaluate the efficacy of a neoadjuvant chemotherapy treatment regimen containing cisplatin, doxorubicin, and cyclophosphamide (CAP regimen) followed by adjuvant docetaxel in a population of locally advanced (stage IIB/III) TNBC.

In this prospective phase II trial of predominantly stage III patients with TNBC treated with neoadjuvant CAP, 19.5% of patients reached a pCR. After a median follow-up of 4 years, the proportion of patients free of recurrence or death at 3 years was of 51.8%, while the proportion of patients alive at 3 years was of 55.5%. The majority of patients received six cycles of CAP (82.9%) with no new safety signals.

The neoadjuvant treatment of breast cancer has a triple objective: to improve surgical options (ressectability and breast conservation techniques), determine in vivo tumor sensitivity to treatment and improve long-term survival outcomes (20). Thus, neoadjuvant chemotherapy is the standard of care for large tumors and pCR an informative biomarker in this setting, given the likelihood to positively impact all the objectives above (7, 21, 22). Despite being less common, pCR is also documented in large tumors and seems as informative, in terms of long-term outcomes, as for smaller tumors (22). In this trial of patients with predominantly stage III TNBC (37/41 patients with stage III tumors), the use of neoadjuvant cisplatin in combination with doxorubicin and cyclophosphamide was associated with a pCR rate of 19.5% (8/41 patients). Despite the large tumor volume at diagnosis, this is a substantial response. Moreover, it is of note that the majority of patients went on to successfully receive surgery with curative intent. These response outcomes are in line with older trials that tested several non-platinum-based regimens in patients with unselected (not exclusively TNBC) and large breast cancers (with 4 cm or more): in this setting, pCR in the breast ranged from 17 to 29% (23–25). Likewise, more recent studies comparing the efficacy of several contemporaneous non-platinum-based regimens, and including tumors as small as 1–2 cm, reported pCR rates ranging from 6 to 26% (26–31). Of note, only in recent times were pCR definitions standardized (22); thus, estimates from previous studies might misrepresent what would be the current pCR rates of such studies.

In TNBC, and more specifically in the basal-like intrinsic subtype, several preclinical studies showed a particular sensitivity to chemotherapy, and especially to DNA-damaging drugs, such as alkylating agents. In this setting, several trials tested the role of neoadjuvant platinum salts in TNBC, obtaining pCR rates ranging from 21 to 61% (8, 13–15, 32, 33). Moreover, recent updates of survival results of GeparSixto trial further showed an 44% improvement in event-free survival (EFS; HR 0.56, 95% CI 0.33–0.96), which translated into an absolute 9.7% improvement in the risk of recurrence at 3 years (76.1 vs. 85.8% for carboplatin-treated patients) (34). A less pronounced, and statistically non-significant, effect of neoadjuvant carboplatin was obtained in the CALGB 40603 trial (EFS HR 0.84, 95% CI 0.58–1.22; 3 years EFS 71 vs. 76% for the carboplatin arm) (35). It is noteworthy that neither GeparSixto nor CALGB 40603 were powered to show a survival difference. In this current trial, after neoadjuvant CAP, the majority of patients received adjuvant docetaxel, and the estimated DFS and OS at 3 years was 51.5 and 55.5%, respectively. These results compare poorly with the trials above; however, they reflect a different patient population, with 32% stage III tumors in the CALGB 40603 trial while 90.2% in the current trial and 59% cN0 in the GeparSixto trial while 22.0% in the current study. Ongoing trials in TNBC, as the NRG-BR003 (NCT02488967) of adjuvant doxorubicin and cyclophosphamide followed by paclitaxel plus or minus carboplatin, and ECOG1131 (NCT02445391) of post-neoadjuvant cisplatin or carboplatin vs. capecitabine in patients with residual disease might further contribute to clarify the role of platinum agents in TNBC.

BRCA status is being studied as a predictive biomarker of response to platinum salts, based on the preclinical evidence that tumors harboring BRCA deficiency are homologous recombination deficient and thus more sensitive to DNA damaging agents, as platinum salts. However, the predictive role of BRCA status is still not clear in this setting. A sub-analysis of GeparSixto showed that response to carboplatin and survival outcomes were independent from germline BRCA (gBRCA) status (34). When comparing pCR rates in patients receiving or not carboplatin, in gBRCA wild-type patients these were of 33.1 vs. 50.8% (OR 2.09, 95% CI 1.24–3.53), respectively, and for gBRCA mutant these were 50.0 vs. 61.5% (OR 1.60, 95% CI 0.52–4.93), respectively. This shows that gBRCA-mutated patients seem to respond more to chemotherapy, but not specifically to addition of carboplatin. Based on these data and the fact that cases of gBRCA mutation tend to cluster in families with known FHBOC, we tested how known FHBOC impacted survival outcomes. Our data show that patients with FHBOC might have an improved recurrence and survival outcomes. A lack of comparator in our trial hinders the possibility to extricate the predictive and/or prognostic information of this particular family history.

The CAP regimen was first developed in ovarian cancer where its safety profile is better established (17–19). In the setting of breast cancer, the concurrent administration of these agents is mostly untested. Such treatment intensification was used as a strategy to maximize response in a group of patients with a large volume of disease. Despite such therapy intensification, no unexpected new safety signals were documented. Nausea (95.3%), vomiting (79.1%), alopecia (79.1%), and fatigue (74.4%) of any grade were the four most common AEs. The tolerability of CAP compares unfavorably with other platinum-based neoadjuvant regimens for breast cancer of less intensified (14) and of more intensified administration of chemotherapy (8). Of note, at the time of the trial design and patient accrual, current drugs for highly emetic regimens, including neurokinin 1 receptor antagonist and olanzapine, were not available. In addition, 3 patients (7.0%) presented febrile neutropenia, of which 1 as a SAE, and paresthesia was present in four patients (9.3%), all grade 1 or 2. The safety profile is comparable to other cisplatin-based neoadjuvant regimens (8, 14). Whether platinum salts should be provided to all, a selected subpopulation, or none of the TNBC; if they should be provided in combination with anthracyclines, taxanes or both; the preferred platinum agent, and the correct dosing/scheduling of these agents are still a matter of active research. These are some of the questions that will continue to shape the coming years of clinical research in this topic.

Furthermore, besides tackling the unmet need of improving outcomes of patients with LABC, a substantial effort should also be allocated to reduce the incidence of such cases. In Brazil, approximately 71% of patients present with stage II/III breast cancer, which contrasts with most westerns countries where between 45 and 55% of patients present with stage II/III tumors (2, 36). Generating the conditions to grant the access to clinical care in settings as Brazil is, therefore, of cornerstone importance.

Despite the interesting findings of this study, it presents several limitations. It is a small, single center trial with no comparison group, which limits the interpretation of the findings. Despite the collection of FHBOC, the specific role of BRCA status and overall homologous recombination deficiency was not addressed in this study. Also, the study was not powered to assess the survival impact of the CAP regimen. Moreover, the 5% pCR rate benchmark used as the null hypothesis for the first stage of the trial, and thus as the rule for early termination, is lower than that usually found in unselected TNBC. However, such cutoff reflects a cohort of predominantly stage III disease, the fact that taxanes were only provided in the adjuvant setting, and previous similar results at the local institution (unpublished data). Finally, as the tumor assessment at diagnosis was performed clinically, the accurate tumor size at diagnosis might have been under or over-estimated.

Cisplatin in association with doxorubicin and cyclophosphamide was associated with a pCR rate (ypT0ypN0) of 19.5% in a cohort of patients with predominantly stage III tumors. Family history of breast and ovarian cancer did not stratify responders in terms of pCR, but identified patients that might do better in terms of DFS and OS. The tolerability profile of this combination poses clinical challenges to its general use in clinical practice.

This study was carried out in accordance with the recommendations of Comissão Nacional de Ética em Pesquisa with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Comissão Nacional de Ética em Pesquisa.

AF: study procedures, data analysis, data interpretation, and manuscript writing. OM-F and RS: study procedures, data interpretation, and manuscript writing. JB: study design, study procedures, data interpretation, and manuscript writing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.