Abstract

Introduction

Triple-negative breast cancer (TNBC) accounts for approximately 15% of all breast cancers and represents a great clinical challenge in the clinic, since it is associated with a larger rate of recurrence and a poorer survival [1]. TNBC is characterized by the absence of hormonal receptors and no amplification of human epidermal growth factor receptor-2 (HER2) gene [2].In contrast with other subtypes, systemic treatments for early TNBC have been restricted to traditional chemotherapy regimens for decades.

For patients with early stage TNBC, the use of NACT has become a standard approach [3], despite its impact on the long-term outcomes being controversial [4]. The main aims of neoadjuvant chemotherapy (NACT) are to reduce the extent of surgery, to attain the good prognostic impact of pathologic complete response (pCR) and to guide adjuvant therapy according to the response. Approximately 30–40% of all TNBC patients achieve a pCR after standard neoadjuvant regimens including anthracycline, taxane and cyclophosphamide [2, 5]. TNBC patients who achieve pCR after NACT have shown a significantly reduced risk of relapse and death, compared with patients with residual disease – consequently, it is widely accepted that achieving pCR has a strong favorable prognostic value [6]. Moreover, pCR is associated with lower rates of systemic and local recurrence, as well as a predictor of excellent survival regardless of tumor subtype [6, 7]. Hence, optimization of NACT regimens with the aim of increasing pCR rates has been considered a promising approach for improving prognosis in TNBC.

Approximately 75% of breast cancers containing germline mutations in BRCA genes (gBRCA) show a triple negative phenotype, with BRCA1 dysfunction frequently being one of the main drivers [8]. Among all of patients with TNBC, 10–15% of patients have gBRCA mutations [9]. Breast cancers with germline BRCA1 or BRCA2 pathogenic or likely pathogenic variants and biallelic inactivation show evidence of deficiency in homologous recombination repair [10, 11].

The loss of BRCA function may turn these tumors particularly sensitive to DNA damaging agents, including platinum agents and poly [ADP- ribose] polymerase inhibitors (PARPi). In patients with gBRCA mutations, PARPi have proved to be an effective treatment option in the metastatic setting [12–14] and are currently they are being explored in the early setting of the disease [15]. Platinum agents (i.e. carboplatin and cisplatin) are cytotoxic DNA damaging compounds leading to DNA strand breaks; this mechanism of action is especially active in cancer cells with DNA repair deficiency such as those harboring deleterious mutations in BRCA genes. In TNBC patients, platinum-based NACT is associated with significantly increased pCR rate [16]. These agents have shown activity in cancers with gBRCA mutation, as BRCA 1/2 proteins have an essential role at repairing the DNA damage [17, 18]. However, the efforts to select a clinically or biologically defined subgroup of patients, who will benefit from the addition of carboplatin, have to date not been conclusive [19]. Several trials demonstrated the effectiveness in the preoperative setting of platinum-based chemotherapy for TNBC patients with gBRCA mutations [18]; although, two randomized clinical studies showed that the addition of platinum to standard neoadjuvant chemotherapy significantly increased pCR rate in TNBC regardless of the presence of BRCA mutation [16]. Nevertheless, BRCA status is considered a predictive factor of response to chemotherapy leading to higher pCR rates and better disease-free survival in the neoadjuvant setting [20–22].

Bevacizumab is a humanized monoclonal antibody that targets the main isoforms of circulating vascular endothelial growth factor (VEGF), resulting in the inhibition of angiogenesis, cell tumor growth, and cell survival. Bevacizumab use has been investigated in both advanced and early-stage breast cancer treatments, showing an increased response rate, mainly in TNBC patients [23]. The treatment of gBRCA mutated breast cancer patients through the use of directed agents for that patient subset is an active area of research.

Since only one third of patients responds to chemotherapy, the identification of novel molecular drivers is crucial for the development of effective targeted treatments. Recently, several clinical trials researching beyond conventional cytotoxic agents showed promising results [24].

To improve the outcome of patients with gBRCA mutated TNBC, several approaches for increasing the efficacy of NACT have been pursued. This systematic review and meta-analysis intend to evaluate the impact of different neoadjuvant treatments in pCR rates in this population, beyond traditional standard chemotherapy.

Methods

We performed the present systematic review according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [25]. The Prospero registration number is CRD42020192946.

Search strategy and selection criteria

The following electronic bibliographic databases were systematically searched: MEDLINE, Web of Science database, Embase and Cochrane CENTRAL. All clinical trials regarding NACT in early BRCA-mutated TNBC that were published from 2001 to 2021 were retrieved, with no language restriction. Abstracts and presentations from the American Society of Clinical Oncology (ASCO), the European Society for Medical Oncology (ESMO) and the San Antonio Breast Cancer Symposium (SABCS) from 2001 to 2021, were also reviewed to identify relevant unpublished studies.

Two investigators (OC and CS) independently searched the databases, using search terms mainly relating to neoadjuvant treatment in BRCA TNBC patients. Specific keywords and free text terms were combined with Boolean operators. The following search phrase was used: (breast OR mammary) AND (cancer OR cancers OR tumor OR neoplasm OR carcinoma) AND (BRCA) AND (neoadjuvant chemotherapy OR induction chemotherapy OR pre-operative chemotherapy) AND (TNBC OR triple-negative OR triple negative OR basal-like OR HER2 negative) AND (pathological complete response OR pCR), without any limits or restrictions. To be eligible, studies had to meet the following criteria: (1) prospective, retrospective or randomized clinical trial in patients with pathogenic BRCA mutated early TNBC; (2) must have enrolled BRCA TNBC patients receiving NACT; (3) must have provided data on pCR. We excluded (1) case reports, reviews, meta-analyses, animal, or in vitro studies; (2); ongoing studies with results not presented nor published at the time of the literature search.

The investigators (OC, CS) independently double-screened and reviewed the list of records retrieved in accordance with the above-mentioned criteria, to identify potentially eligible articles. When discrepant opinions on study selection among investigators occurred, a third author functioned as tiebreaker; when no compromise was reached, all authors were consulted.

The PRISMA flowchart (Fig. 1) summarizes the process of the search strategy for study selection (screening, eligibility, inclusion).

Open in a separate window

Fig. 1

Literature search PRISMA flowchart

Data extraction

The following information was extracted and included: study name, first author, year of publication, study design, stage of disease, number of BRCA mutated TNBC patients enrolled per regimen(s) of NACT or other treatments, number of patients achieving pCR. Toxicity profile, specifically number of patients with grade 3 or 4 adverse events (AEs) was also retrieved when available. A funnel-plot was performed to assess for potential publication bias -Fig. 10.

Open in a separate window

Fig. 10

Funnel plot assessment of publication bias for pCR in patients receiving NACT and platinum-based therapies for early TNBC with gBRCA mutations

Results

A total of 1228 studies was identified through the initial search strategy (Fig. 1). After screening the abstracts and reviewing the full texts, a total of 31 trials involving 619 gBRCA mutated TNBC patients were selected for the final analysis.

Among the 619 gBRCA mutated TNBC patients included in the analysis, 139 patients received cisplatin alone (Table ​(Table1).1). Table ​Table22 shows patients (n=133) who were treated with platin derivatives combined with anthracyclines and taxanes. One single study [28] included 10 patients who were not treated with taxanes, and 18 patients from other study [29]. Fifty-three patients received a combination of standard NACT (anthracycline, cyclophosphamide, taxanes) with carboplatin (Table ​(Table33).

Table 3

Characteristics of studies included based on treatment regimens: standard chemotherapy regimen (anthracycline, cyclophosphamide, taxanes) and carboplatin

	Affiliation	Type of study	Stage of disease	Nº of BRCA1/2 mutated TNBC patients	pCR	Neoadjuvant treatment
Fontaine 2019 [26]	Belgium	RCT phase II	II-III	9	7	Cb-P+E-C
Sella 2018 [35]	Israel	Clinical trial	I-III	14	9	Cb-P-Dox-C
Walsh 2019 [36]	Irland	Retrospective	I-III	6	4	Cb-P+Dox-C
Loib 2018 BrighTNess [37]	USA	RCT phase III	II-III	24	12	Cb-P- Dox-C

Open in a separate window

pCR Pathological complete response, Cb Carboplatin, Dox Doxorubicin, P Paclitaxel, E Epirubicin, C Cyclophosphamide

Table ​Table44 presents selected trials with patients treated with carboplatin and taxanes (n=108) and Table ​Table55 present one single study [38] in which patients were treated with carboplatin, iniparib and gemcitabine. Eighty-three patients received standard chemotherapy, carboplatin and a PARPi (Table ​(Table6)6) and 19 patients received a PARPi alone (Table ​(Table7).7). Table ​Table88 describes cases (n=65) who were treated with an anti-VEGF agent associated with standard chemotherapy and carboplatin. Only three patients were treated with eribuline and carboplatin (Table ​(Table99).

Table 6

Characteristics of studies included based on treatment regiments: carboplatin+standard NACT+PARPi

	Affiliation	Type of study	Stage of disease	Nº of BRCA1/2 mutated TNBC patients	pCR	Neoadjuvant treatment
Severson 2017 [46]	Holand	Multicenter phase II trial	II-III	32	17	Cb-P–V+Dox-C
Loib 2018 BrighTNess [37]	USA	Multicenter, RCT phase III trial	II-III	46	26	Cb-P–V+Dox-C
Litton 2017 [47]	USA	pilot trial,	I-III	5	3	Cb-P–T-Dox-C

Open in a separate window

pCR Pathological complete response, Cb Carboplatin, Dox Doxorubicin, C Cyclophosphamide, P Paclitaxel, V Veliparib, T Talazoparib

Table 8

Characteristics of studies included based on treatment regiments standard NACT+anti-VEGF antibody

	Affiliation	Type of study	Stage of disease	Median age	Nº of BRCA1/2 mutated TNBC patients	pCR	Neoadjuvant treatment
Hahnen 2017 Gepar Sixto [49]	Germany	Phase II RCT	II-III	48	26	17	Cb-P-Dox-Beva
Fasching 2018 Gepar Quinto [21]	Germany	Phase III RCT	I-III	48	39	23	E-C+D-Beva

Open in a separate window

pCR Pathological complete response, Cb Carboplatin, Dox Doxorubicin, C Cyclophosphamide, P Paclitaxel, D Docetaxel, E Epirubicin, Beva Bevacizumab

Proportion of pCR achieved

In gBRCA mutated TNBC patients who received cisplatin in monotherapy, the proportion of patients who achieved pCR was 0.53 (95%CI [0.30, 0.76]) (Fig. 2). When treatment was a combination of standard chemotherapy and platin derivatives the proportion of pCR increased to 0.62 (95%CI [0.48, 0.76] (Fig. 3). Similarly, the group who received carboplatin and taxane achieved a proportion of pCR of 0.63 (95%CI [0.47, 0.79]) (Fig. 4).

Open in a separate window

Fig. 2

Forest plot- Cisplatin in monotherapy

Open in a separate window

Fig. 3

Forest plot- Standard chemotherapy and carboplatin

Open in a separate window

Fig. 4

Forest plot- Carboplatin and taxanes

The group of patients treated with platin derivatives, anthracyclines±taxanes achieved the highest proportion of pCR, 0.66 (95%CI [0.57, 0.76]) (Fig. 5).

Open in a separate window

Fig. 5

Forest Plot- Carboplatin+anthracyclines±taxanes

With respect to gBRCA mutated TNBC patients treated with PARPi alone pCR achievement of was seen in a proportion of 0.55 (95%CI [0.30, 0.81]) (Fig. 6). When standard chemotherapy and platin derivatives were combined with PARPi the proportion of pCR did not change, 0.55 (95%CI [0.45, 0.66]) (Fig. 7).

Open in a separate window

Fig. 6

Forest Plot- PARP inhibitor alone

Open in a separate window

Fig. 7

Forest Plot- Carboplatin+Standard NACT+PARPi

The group of patients treated with taxane, anthracycline and anti-VEGF (bevacizumab) achieved a proportion of pCR of 0.62 (95%CI [0.50, 0.73]) (Fig. 8) although one study also included carboplatin [49], and another cyclophosphamide in the treatment regimens [21].

Open in a separate window

Fig. 8

Forest Plot- Carboplatin+Standard NACT+anti VEGF antibody

Only one study evaluated the association of carboplatin and eribulin, which achieved pCR proportion in two-thirds of patients [50].

Figure 9 displays the proportion of pCR achieved with different treatment regimens and the corresponding number of patients. The largest group was treated with cisplatin in monotherapy and achieved the lowest proportion of pCR achievement. On the other hand, the highest pCR was achieved in the group treated with platin derivates, anthracyclines±taxanes which included the second highest number of patients.

Open in a separate window

Fig. 9

Bubbleplot graphic showing the proportion of different treatments in pCR achievement. The size of a bubble represents the number of patients included in each group

The symmetric funnel plot for this meta-analysis shows an additional indicator of the absence of publication bias and study heterogeneity (Fig. 10).

Safety outcomes

Hematological effects

Several studies reported grade 3 and 4 hematological AE (i.e. neutropenia, anemia, and thrombocytopenia) (Table ​(Table10).10). However, most of them did not report adverse effects according to BRCA status. The most common hematological adverse effect was neutropenia. This was very significant in patients treated with carboplatin, standard NACT and anti-VEGF agent, with an incidence of 76.6%. On the other hand, groups treated with PARPi or cisplatin alone presented the lowest incidences of neutropenia, 5.8% and 3.6% respectively, while the group treated with carboplatin and eribulin revealed an incidence of 60%. Concerning anemia, the highest incidence was reported in the group treated with carboplatin, standard NACT and PARPi (25.1%).

Table 10

Incidence of hematological adverse effects with different neoadjuvant treatment regimens, regardless of BRCA status

	Cisplatin alone	Carboplatin+anthracycline+taxane	Carboplatin+taxane	Carboplatin+Gemcitabine+PARPi	Carboplatin+NACT+PARPi	Carboplatin+NACT+anti-VEGF antibody	PARPi	Carboplatin+Eribulin
n	196	106	239	93	398	1246	52	30
Adverse effect
Neutropenia	3.6%	50.9%	10.5%	45.2%	59.0%	76.6%	5.8%	60.0%
Anemia	1.0%	20.8%	4.2%	8.6%	25.1%	5.0%	15.4%	23.3%
Febrile neutropenia	0.5%	18.9%	3.3%		1.5%	12.4%
Thrombocytopenia		12.3%	5.4%	6.5%	13.1%	5.0%	1.9%	20.0%
Leukopenia		12.3%	0.8%		4.5%

Open in a separate window

Non-hematological effects

Table ​Table1111 presents a summary of reported non-hematological AEs. The group of carboplatin, standard NACT and anti-VEGF agent showed the highest incidence of gastrointestinal adverse effects (27,7%) cardiac disorders (6,7%), renal and urinary (4,8%) and skin and subcutaneous tissues AEs (7,1%). The group treated with carboplatin, gemcitabine and PARPi also presented considerable gastrointestinal symptoms, in up to 24,7% of cases.

Table 11

Incidence of non-hematological adverse effects with different neoadjuvant treatment regimens

	Cisplatin alone	Carboplatin+anthracycline+taxane	Carboplatin+taxane	Carboplatin+Gemcitabine+PARPi	Carboplatin+NACT+PARPi	Carboplatin+NACT+anti VEGF antibody	PARPi	Carboplatin+Eribulin
n	196	106	239	93	398	1246	52	30
Adverse effect
Gastrointestinal disorders	4.6%	2.8%	8.4%	24.7%	6.5%	27.7%	0.0%	0.0%
Cardiac disorders	1.5%	0.0%	0.4%	1.1%	0.8%	6.7%	0.0%	0.0%
Nervous system disorders	0,.0%	2.8%	1.3%	2.2%	0.8%	1.5%	0.0%	0.0%
Renal and urinary disorders	1,.0%	0.0%	0.4%	1.1%	0.0%	4.8%	0.0%	3.3%
Skin and subcutaneous tissue disorders	1.0%	0.0%	0.8%	0.0%	0.0%	7.1%	0.0%	0.0%

Open in a separate window

Discussion

The goal of this systematic review was to assess the proportion of pCR in gBRCA TNBC patients when neoadjuvant treatments regimens other than standard ones were used.

A lot of efforts has been done to identify predictive markers for the use of platinum, driven by the hypothesis that tumors with deficient homologous recombination, such as those with gBRCA1/2 mutations, may be better targeted by carboplatin due to their inability to repair double-strand DNA breaks induced by platinum salts. However, its efficacy for breast cancer with BRCA germline mutations remains inconclusive.

Our results point to a significant role of standard NACT (anthracyclines, taxanes and cyclophosphamide) in this setting, as the combination of carboplatin with standard regimens yielded a proportion of pCR achievement of 0.62 (95%CI 0.48–0.76), higher than in patients treated with cisplatin alone [0.53 (95%CI 0.30–0.76); risk ratio 1.234 (95% CI 0.936–1.672)].

Accordingly, in our meta-analysis, patients treated with platin derivatives plus an anthracycline with or without a taxane (two studies without taxanes) achieved the highest proportion of pCR, 0.66 (95%CI [0.57, 0.76]), closely followed by the group who received carboplatin and taxane, in which a proportion of pCR of 0.63 (95%CI [0.47, 0.79]) was achieved. Interestingly, analysis of pCR with anthracyclines and taxanes in the presence of platin derivates favored anthracyclines, although hematological AEs increased with these agents.

Previously published meta-analysis that assessed the addition of platinum to standard NACT found an improvement of pCR rates for patients with BRCA mutations, although this was not a statistically significant [51,16]. Similar results were obtained in our previous meta-analysis, which revealed an increased pCR rate in BRCA mutation carriers (58.4%) compared with non-carriers (50.7%), but with no statistical significance [19].

The lowest proportion of pCR rate in our study (0.53 (95%CI [0.30, 0.76]) was found in the group of 139 patients treated with cisplatin in monotherapy.

Besides platin agents, other neoadjuvant treatments were reported in the trials included in this meta-analysis. PARP enzymes play a major part in DNA repair mechanisms and inhibition of PARP activity leads to the accumulation of double-strand DNA breaks. These breaks are normally repaired by double-strand homologous recombination pathways that include the tumor-suppressor proteins BRCA1 and BRCA2. Thus, gBRCA mutated TNBC as well as the BRCAness phenotype are in theory particularly vulnerable to PARPi [52].

In our study, when PARPi (talazoparib or olaparib) were used in monotherapy pCR was the same (0.55 (95% CI [0.30, 0.81])) as when added to standard chemotherapy and carboplatin (0.55 (95% CI [0.54, 0.66])), only with a much lower incidence of hematologic toxicity.

However, when PARPi were added to standard chemotherapy and carboplatin, proportion of pCR was lower than that of treatment with standard chemotherapy and carboplatin in the absence of PARPi (0.62 (95%CI [0.48, 0.76])). These results point to a neutral to non-beneficial effect of PARPi in this setting. This is surprising as it would be expected that the addition of PARPi would increase pCR rate when compared to those treated with standard chemotherapy and carboplatin. This has been shown by a recent study which found significantly longer survival free of invasive or distant disease when using olaparib as adjuvant therapy after neoadjuvant or adjuvant chemotherapy and local therapy in early breast cancer patients with BRCA1 or BRCA2 germline pathogenic variant [53]. Likewise, in the I-SPY2 phase 2 trial, the addition of veliparib and carboplatin to standard NACT improved pCR from 26% in the control arm to 51% in the veliparib–carboplatin group of TNBC patients [54].

The VEGF pathway plays a key role in the pathophysiology of TNBC. However, in our study, the addition of bevacizumab to standard chemotherapy with platin derivatives did not yield any benefits since a similar proportion of pCR achievement was obtained in both groups 0.62 (95%CI [0.50–0.73]). Moreover, patients in this group reported the higher incidence of neutropenia (over 76%).

The evaluation of pCR is of extreme importance. However, the real impact of this outcome in long term clinical results is not yet clear. In this study we tried to consider other outcomes but few of the included studies reported long-term outcomes in relation to BRCA status. The vast majority did not discriminate between subgroups and reported outcomes like DFS or OS for the entire group of TNBC patients. Only two studies, GeparSixto [49] and Fasching et al. [21], separately analyzed BRCA mutated TNBC patients and found that pCR was a strong predictor of DFS for patients without BRCA, but not for patients with BRCA mutations. Nevertheless, with regard to prognosis, patients with a BRCA pathogenic variant had a significantly better DFS.

It is important to point out that few trials were sufficiently powered enough to assess long-term outcomes in the TNBC gBRCA mutated group. Hence, the question of clinical utility of different treatment approaches in this subgroup remains unanswered and further research is necessary.

A recent publication exploring safety issues in the neoadjuvant setting concludes that gBRCA1/2 mutated patients show a higher risk of hematological toxicity when treated with regimens including a taxane [55]. On the contrary, our study demonstrates higher adverse hematological AEs with the addition of anthracyclines, PARPi and anti-VEGF to a standard regimen with platin. This difference may be related to the lack of distinction of AEs according to the BRCA1/2 status.

Our study presents several limitations. Major limitations are related to the small number of patients with gBRCA TNBC included in the different trials and heterogeneity between trials (related to study design, drugs and doses of treatment regimens). Nonetheless, gBRCA mutated TNBC patients are rarely distinguished in trials and such approach is considered the only way to obtain conclusions.

Almost 20% of breast cancer patients share histological features and clinical outcomes with BRCA1/2 related cancers without detectable gBRCA1/2 mutations, a phenotype defined as BRCAness. Beyond gBRCA mutations, somatic BRCA mutation and BRCA silencing through promotor hypermethylation or alterations affecting other genes related to homologous recombination [10] that can mimic the BRCAness state. Importantly, BRCA1 methylated and gBRCA1 mutated TNBCs share gene expression and immune profiles and seem to have a similar outcome after adjuvant chemotherapy [56]. Consequently, another limitation of our study was to include only gBRCA mutations and not all cases with the BRCAness phenotype.

To our knowledge this is the first study that gathers information on gBRCA mutation TNBC patients, a subgroup with many singularities often not separately analyzed in published trials. Our assessment of neoadjuvant treatments in this distinct group of TNBC revealed clinically relevant conclusions with possible impact on treatment options. It is also noteworthy that this is the first study in this subset of patients of such a wide range of treatments beyond conventional chemotherapy.

Conclusions

This study showed that patients with gBRCA mutated TNBC patients treated with cisplatin in monotherapy in the neoadjuvant setting present a lower pCR when compared with standard chemotherapy combined with platin derivatives, strengthening the role of standard chemotherapy. Likewise, the addition of PARPi to standard chemotherapy and carboplatin decreased the proportion of pCR denoting no contribution of PARPi in this setting and favoring the role of standard chemotherapy and platin derivates. The highest proportion of pCR was found with the combination of platin derivates and anthracyclines±taxanes.

Acknowledgements

Authors’ contributions

OC conceived the literature review and the design of the review. OC and CS undertook all the literature searches and data extraction. OC, CS and FC assumed the statistical analysis. OC, CS and FC drafted the manuscript and all authors read, edited and approved the final manuscript.

Funding

This work received no funding.

Availability of data and materials

Not applicable.

Declarations

Footnotes

References