Platinum-based neoadjuvant chemotherapy in triple-negative breast cancer: An updated systematic review and meta-analysis

Abstract

Background:

Triple-negative breast cancer (TNBC) is a highly aggressive subtype of breast cancer that is characterized by its association with shortened survival durations and a heightened likelihood of recurrence. Platinum-based chemotherapy has been demonstrated to increase the rate of pathological complete response (pCR), yet its influence on long-term survival outcomes remains unclear. This meta-analysis aims to clarify the activity, efficacy, and safety of platinum-based regimens in this patient population.

Methods:

We conducted a systematic search of PubMed and Web of Science up to October 13, 2024, to identify randomized controlled trials (RCTs) comparing platinum-based neoadjuvant chemotherapy (NACT) to platinum-free regimens in TNBC patients. Using random effects models, we calculated pooled odds ratios and hazard ratios with 95% confidence intervals for pCR (defined as ypT0/is pN0), event-free survival (EFS), overall survival, distant disease-free survival, invasive disease-free survival, and grade 3 and 4 adverse events.

Results:

A total of 12 RCTs involving 2650 patients were included in the analysis. Overall, platinum-based NACT significantly increased the pCR rate from 34.6% to 51.3%. This association remained significant when restricting the analysis to 7 RCTs (N = 1645) that utilized the same standard regimen of weekly paclitaxel (with or without carboplatin) followed by anthracycline and cyclophosphamide. Regarding survival outcomes, various RCTs reported different metrics: EFS, overall survival, distant disease-free survival, and invasive disease-free survival showed no significant differences.

Conclusions:

Our meta-analysis showed that platinum-based NACT was associated with significantly increased pCR and EFS rates in TNBC patients. Although it entails an increased risk of grade 3/4 hematological toxicity, the risk is still manageable. These findings suggest that the addition of a platinum agent to standard neoadjuvant anthracycline- and taxane-based chemotherapy may be considered an option in TNBC patients.

1. Introduction

Triple-negative breast cancer (TNBC) is characterized by the absence of estrogen receptor, progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression.^[1] When considered as a whole, this heterogeneous group of tumors exhibits the highest rate of distant metastasis and the lowest overall survival (OS) among all breast cancer subtypes.^[2] Despite surgery and adjuvant therapies, approximately half of primary TNBC cases confined to the breast and lymph nodes will recur at distant sites within 5 years, with a strong predilection for metastasis to visceral organs and the central nervous system.^[3] Currently, the systemic treatment of metastatic TNBC is largely confined to chemotherapy drugs, with successive regimens showing progressively decreasing effectiveness. Despite a substantial understanding of its molecular landscape, no biologically targeted therapies have yet proven efficacious for this subtype. Consequently, TNBC poses a significant challenge for clinicians due to its aggressive nature and the lack of effective treatment options comparable to endocrine therapy for estrogen receptor-positive breast cancer or anti-HER2 agents for HER2-positive breast cancer.

Given that there are no differences in survival between adjuvant and neoadjuvant treatment settings, neoadjuvant chemotherapy (NACT) is now regarded as the standard approach for managing high-risk TNBC. It aims to reduce the tumor burden and assess chemotherapy efficacy prior to surgical resection.^[4] NACT offers several distinct advantages, including: the potential to downstage the primary breast tumor, which may allow for a less extensive surgery (e.g., breast-conserving surgery instead of mastectomy); a reduction in the extent of axillary node resection (e.g., sentinel node biopsy instead of axillary lymph node dissection); the opportunity to consult with plastic surgeons and genetic counselors prior to surgery; and most importantly, the ability to assess tumor response to treatment, prognosis, and the potential need for additional or adjuvant therapies. A complete pathological response, characterized by the complete disappearance of invasive cancer post-NACT, has been strongly associated with favorable outcomes.^[5,6]

The phenomenon of higher pathological complete response (pCR) rates in TNBC following NACT juxtaposed with worse survival outcomes compared to non-TNBC tumors is referred to as the “triple-negative paradox.”^[7] Platinum compounds, including cisplatin, carboplatin, and oxaliplatin, have been pivotal in oncology for decades, primarily due to their mechanism of action, which involves the formation of DNA-protein cross-links, inhibiting DNA synthesis and function, and ultimately leading to cell apoptosis.^[8] The enhancement of neoadjuvant treatment through the integration of platinum-based regimens has been a focal point in recent years, aimed at increasing pCR rates in TNBC.

The combination of platinum with conventional taxane and anthracycline regimens has been shown to elevate pCR rates from roughly 35% to over 50% in TNBC patients.^[9] However, the benefits of platinum addition to NACT must be weighed against increased toxicity profiles, particularly hematological events. Some literature advocates against the routine addition of platinum drugs to standard NACT for TNBC,^[10] highlighting the need for a balanced approach that considers the potential toxicity that could limit timely and effective therapy.^[8]

By critically evaluating the existing data, we endeavor to elucidate the role of platinum-based NACT in TNBC patients, thereby contributing to the ongoing discourse on optimizing treatment strategies for this challenging subtype of breast cancer. Through this systematic review, we seek to address the controversies surrounding the use of platinum in NACT for TNBC, considering both the positive and negative aspects, and aiming to inform clinical decision-making with evidence-based insights.

2. Methods

Ethics approval is not necessary. This is a meta-analysis; this study only collected data that had already been publicly published.

This systematic review and meta-analysis synthesized quantitative evidence from randomized controlled trials (RCTs) to evaluate the activity, efficacy, and safety of platinum-based NACT (experimental arm) compared to platinum-free NACT (control arm) in TNBC patients. The methodology was meticulously designed to ensure the robustness and reliability of findings.

2.1. Search strategy and study identification

A comprehensive literature search was conducted in the PubMed and Web of Science databases without language or date restrictions, concluding on October 13, 2024. The search strategy incorporated a combination of specific keywords and free-text terms using Boolean operators to identify relevant studies. Keywords included “breast cancer,” “platinum,” “carboplatin,” “cisplatin,” “neoadjuvant,” and “chemotherapy.” A detailed search expression is provided in the supplementary materials (Supplemental Digital Content, https://links.lww.com/MD/Q310). In addition to the database searches, cross-referencing of relevant articles was employed to capture all potentially pertinent records. Two independent reviewers (JW and XY) executed the systematic literature search, with discrepancies resolved through consultation with a third reviewer (LP).

2.2. Selection criteria and data extraction

For eligibility in the meta-analysis, studies had to fulfill the following criteria: phase II or III RCTs that included TNBC patients receiving platinum-based NACT in the experimental arm versus platinum-free NACT in the control arm. Only studies reporting pCR rates in both arms to estimate odds ratio (OR) and 95% confidence interval (CI) were included. Studies were excluded if they were non-RCTs, investigated platinum-based NACT in breast cancer subtypes other than TNBC, or were ongoing with results unavailable at the time of the literature search. From each included RCT, the following variables were extracted: study name, population characteristics, intervention details, and outcomes related to pCR rates.

2.3. Study objectives

The primary focus is on the absence of residual invasive tumors in the breast or axilla posttreatment. The analysis encompasses 4 main categories: all RCTs regardless of chemotherapy backbone; RCTs where anthracycline- and taxane-based NACT was utilized in both treatment arms; RCTs where the same NACT backbone, with or without a platinum agent, was administered to both groups; RCTs employing the identical standard NACT regimen of weekly paclitaxel (with or without a platinum agent) followed by anthracycline plus cyclophosphamide in both arms.

Secondary objectives include assessing the activity, efficacy, and safety of platinum-based versus platinum-free NACT in TNBC patients concerning overall response rate (ORR), event-free survival (EFS), OS, distant disease-free survival (DDFS), invasive disease-free survival (iDFS), stable disease (SD), progressive disease (PD), and grade 3 and 4 adverse events (AEs) such as neutropenia, anemia, neuropathy, leukopenia, and thrombocytopenia.

2.4. Statistical analysis

The ORs and 95% CIs for the effects of platinum-based and platinum-free NACT on pCR, ORR, and grade 3 and 4 AEs were calculated. An OR >1 suggests higher rates of pCR, ORR, and grade 3 and 4 AEs in the platinum-based NACT group, while an OR <1 indicates lower rates in this group. The Mantel–Haenszel method was employed to obtain a fixed effects model of the pooled ORs,^[11] and standard checks for the homogeneity assumption were performed.^[12] In cases of significant heterogeneity among the trials, the DerSimonian and Laird method was used to compute the pooled estimate of the ORs via a random effects model.^[13] The Higgins I² index was calculated to quantify the degree of inconsistency in study results.^[14] Publication bias was assessed through visual inspection of a funnel plot for study size against treatment effect^[15] and Harbord asymmetry test.^[16] Statistical significance is considered if the 95% CI does not include 1.0, with a P-value < .05 (2-sided). Sensitivity analysis, aimed at determining the stability of the merged OR estimates, was conducted by recalculating the merged OR estimates after excluding each individual study. All statistical analyses and forest plot generation were performed using Revman software version 5.4 (London, United Kingdom).

3. Results

After a comprehensive systematic search, which initially identified 658 records, we excluded 642 irrelevant studies. This left us with 16 potentially eligible RCTs. Among these, 4 were updates of previously published studies, resulting in 12 unique RCTs being included in the final meta-analysis (Fig. 1).

Figure 1.

The PRISMA flow chart summarizing the process for the identification of eligible randomized controlled trials. *Consider, if feasible to do so, reporting the number of records identified from each database or register searched (rather than the total number across all databases/registers). **If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools. From Page et al.^[17] For more information, visit: http://www.prisma-statement.org/.

This meta-analysis encompassed a total of 2650 TNBC patients, with 1315 (49.6%) receiving platinum-based chemotherapy and 1335 (50.4%) receiving platinum-free chemotherapy. The primary characteristics of the included RCTs are summarized in Table 1. Among these RCTs, 10^[18–27] utilized carboplatin as the platinum agent, while the remaining 2^[28,29] employed lobaplatin. In 6^{[18–21,26,27]} RCTs (n = 1868 patients), anthracycline- and taxane-based NACT was administered in both groups. The same chemotherapy backbone, either with or without platinum, was used in 7 of the included RCTs (n = 1645 patients).^{[18,19,21,22,25,26,28]} Additionally, the standard chemotherapy regimen of weekly paclitaxel followed by anthracycline plus cyclophosphamide, with or without carboplatin, was employed in 3 RCTs (n = 826 patients).^[18,19,26]

Table 1.

Main characteristics of the randomized controlled trials included in the present meta-analysis.

Study	Study design	Primary end point	Secondary end points	Treatment arms	TNBC patients, n
BrighTNess[18]	Phase III	ypT0/is N0	EFS, OS, clinical response rate, toxicity	P + Cb → AC	160
				P → AC	158
CALGB 40603[19]	Phase II	ypT0/is N0	ypT0/is pN0, safety, RFS, OS	P + Cb ± Bev → ddAC	221
				P ± Bev → ddAC	212
GeparOcto-dGBG 84[20]	Phase III	ypT0/is pN0	Toxicity, DFS, OS	PDoxCb	203
				DdEPC	200
GeparSixto; GBG 66[21]	Phase II	ypT0/is ypN0	ypT0/is pN0, clinical response rate, safety, efficacy	P + Dox + Bev + Cb	158
				P + Dox + Bev	157
NACATRINE[22]	Phase II	ypT0ypN0	iDFS, OS, toxicity, safety	Dox + C + Cb	203
				Dox + C	200
NCT01276769[23]	Phase II	ypT0/is pN0	ORR, safety, RFS, OS	PCb	44
				EP	43
NeoCART[24]	Phase II	ypT0/is ypN0	EFS, OS, BCS rate, toxicity	DCb	44
				EC-D	44
RSF-2017-00000557[25]	Phase II	ypT0/is ypN0	DDFS, toxicity, safety	AT-Cb	79
				AT	79
UMIN000003355[26]	Phase II	ypT0/is pN0	Clinical response rate, safety, DFS	PCb → CEF	37
				P → CEF	38
WSG-ADAPT-TN[27]	Phase II	ypT0/is pN0	EFS, OS, toxicity	Nab-P + Cb	146
				Nab-P + Gem	178
ChiCTR-TRC-14005019[28]	Phase II	ypT0/is ypN0	DFS, OS, safety	TE + L	99
				TE	101
ChiCTR-TRC-1900023776[29]	Phase II	ypT0/is ypN0	EFS, OS, safety	TL	51
				TEC	52

P + Cb → AC: paclitaxel 80 mg/m² weekly + Cb AUC 6 every 3 weeks for 12 weeks followed by doxorubicin 60 mg/m² plus cyclophosphamide 600 mg/m² every 2 or 3 weeks; P → AC: paclitaxel 80 mg/m² weekly followed by doxorubicin 60 mg/m² plus cyclophosphamide 600 mg/m² every 2 or 3 weeks; P + Cb ± Bev → ddAC: paclitaxel 80 mg/m² once per week for 12 weeks concurrent carboplatin AUC 6 once every 3 weeks for 4 cycles, followed by doxorubicin 60 mg/m² plus cyclophosphamide 600 mg/m² once every 2 weeks for 4 cycles, and/or bevacizumab 10 mg/kg once every 2 weeks for 9 cycles; P ± Bev → ddAC: paclitaxel 80 mg/m² once per week for 12 weeks followed by doxorubicin 60 mg/m² plus cyclophosphamide 600 mg/m² once every 2 weeks for 4 cycles, and/or bevacizumab 10 mg/kg once every 2 weeks for 9 cycles; PDoxCb: paclitaxel 80 mg/m² weekly simultaneously with nonpegylated liposomal doxorubicin 20 mg/m² simultaneously with carboplatin AUC 1.5 weekly for 18 weeks; DdEPC: epirubicin 150 mg/m² every 2 weeks for 3 cycles followed by paclitaxel 225 mg/m² every 2 weeks for 3 cycles followed by cyclophosphamide 2000 mg/m² every 2 weeks for 3 cycles; P + Dox + Bev + Cb: paclitaxel 80 mg/m² plus nonpegylated liposomal doxorubicin 20 mg/m², both once a week for 18 weeks plus bevacizumab 15 mg/kg intravenously every 3 weeks simultaneously plus carboplatin at a dose of 2.0 AUC, once every week for 18 weeks; P + Dox + Bev: paclitaxel 80 mg/m² plus nonpegylated liposomal doxorubicin 20 mg/m², both once a week for 18 weeks plus bevacizumab 15 mg/kg intravenously every 3 weeks simultaneously with all cycles; Dox + C + Cb: doxorubicin (60 mg/m²) plus cyclophosphamide (600 mg/m²) both intravenously (iv) once every 21 days for 4 cycles for all patients; Dox + C: paclitaxel (80 mg/m² iv) once every 7 days for 12 cycles with carboplatin AUC 1.5 (experimental arm) once every 7 days for 12 cycles; PCb: paclitaxel 175 mg/m² on day 1 plus carboplatin AUC 5 on day 2, every 3 weeks for 4 to 6 cycles; EP: epirubicin 75 mg/m² on day 1 and paclitaxel 175 mg/m² on day 2 every 3 weeks for 4 to 6 cycles; DCb: docetaxel (75 mg/m²) administered intravenously every 3 weeks plus carboplain (AUC 6 mg/mL/min, intravenously every 3 weeks for 6 cycles); EC-D: epirubicin (90 mg/m²) plus cyclophosphamude (600 mg/m²), both administered intravenously every 3 weeks for 4 cycles, followed by docetaxel (100 mg/m²) administered intravenously every 3 weeks for 4 cycles; PCb → CEF: 4 cycles of carboplatin AUC 5 every 3 weeks concurrent weekly paclitaxel (days 1, 8, 15) followed by 4 cycles of cyclophosphamide 500 mg/m² plus epirubicin 100 mg/m² plus 50 fluorouracile 500 mg/m² every 3 weeks; P → CEF: weekly paclitaxel (days 1, 8, 15) followed by 4 cycles of cyclophosphamide 500 mg/m² plus epirubicin 100 mg/m² plus 50-fluorouracile 500 mg/m² every 3 weeks; Nab-P + Cb: nab-paclitaxel 125 mg/m² plus carboplatin AUC 2 day 1, 8 every 3 weeks for 12 weeks; Nab-P + Gem: nab-paclitaxel 125 mg/m² plus gemcitabine 1000 mg/m² day 1, 8 every 3 weeks for 12 weeks; TE + L: docetaxel 75 mg/m² as a 1-h intravenous infusion on the first day of every 3-week cycle with concurrent epirubicin at a dose of 80 mg/m² and lobaplatin at a dose of 30 mg/m²; TE: docetaxel 75 mg/m² as a 1-h intravenous infusion on the first day of every 3-week cycle with concurrent epirubicin at a dose of 80 mg/m²; TL: 6-cycle docetaxel 75 mg/m² or albumin-bound paclitaxel 125 mg/m² + lobaplatin 30 mg/m²; TEC: 6-cycle docetaxel 75 mg/m² or albumin-bound paclitaxel 125 mg/m² + epirubicin 75 mg/m² + cyclophophamide 500 mg/m².

BCS = breast-conserving surgery, D-DFS = distant disease-free survival, DFS = disease-free survival, EFS = event-free survival, iDFS = invasive DFS, ORR = objective response rate, OS = overall survival, RFS = relapse-free survival, TNBC = triple-negative breast cancer.

3.1. pCR rates analysis

Overall, in all 12 RCTs, 1137 out of 2650 patients (42.9%) achieved a pCR following neoadjuvant treatment. Specifically, 675 out of 1315 patients (51.3%) in the platinum-based chemotherapy group and 462 out of 1335 patients (34.6%) in the platinum-free chemotherapy group attained pCR (OR 2.00, 95% CI 1.71–2.34, P < .001; I² = 43%, P = .05) (Fig. 2A). Sensitivity analysis revealed that after excluding the GeparOcto-dGBG 84 study (I² = 0, P = .45), the remaining pooled OR estimates remained stable with only minor fluctuations (Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q134). Furthermore, the funnel plot analysis indicated no evidence of publication bias (Fig. 2B).

Figure 2.

(A) Odds ratio for pathological complete response of platinum-based (Platinum) versus platinum-free (Controls) NACT in all included randomized controlled trials (the size of the squares is proportional to the weight of each study). (B) Funnel plot with pseudo 95% confidence limits for the effect of platinum-based NACT estimated from individual studies (horizontal axis) against the study size (vertical axis): publication bias is unlikely as suggested by the symmetric inverted funnel shape. CI = confidence interval, NACT = neoadjuvant chemotherapy, OR = odds ratio.

In 6^{[18–21,26,27]} RCTs utilizing anthracycline- and taxane-based chemotherapy in both arms, 857 out of 1868 patients (45.9%) achieved pCR, with 496 out of 925 patients (53.6%) in the platinum-based group and 361 out of 943 patients (38.3%) in the platinum-free group achieving pCR (OR 1.84, 95% CI 1.53–2.21, P < .001; I² = 65%, P = .01) (Fig. 3A). Detailed sensitivity analyses for these results are presented in Table S2, Supplemental Digital Content, https://links.lww.com/MD/Q134, available at Annals of Oncology online.

Figure 3.

Odds ratios for pathological complete response of platinum-based (Platinum) versus platinum-free (Controls) NACT in the randomized controlled trials using: (A) anthracycline- and taxane-based chemotherapy in both treatment arms; (B) the same chemotherapy backbone in both treatment arms; (C) the same standard NACT regimen in both treatment arms. The size of the squares is proportional to the weight of each study. CI = confidence interval, NACT = neoadjuvant chemotherapy, OR = odds ratio.

In the 7 RCTs employing the same chemotherapy backbone (with or without carboplatin) in both study groups,^{[18,19,21,22,25,26,28]} 718 out of 1645 patients (43.6%) achieved a pCR following neoadjuvant treatment. Specifically, 438 out of 824 patients (53.2%) in the platinum-based chemotherapy group and 280 out of 818 patients (34.2%) in the platinum-free chemotherapy group attained pCR (OR 2.18, 95% CI 1.78–2.66, P < .001; I² = 30%, P = .20) (Fig. 3B). Table S3, Supplemental Digital Content, https://links.lww.com/MD/Q134, available at Annals of Oncology online, presents the sensitivity analysis.

Three RCTs assessed the impact of adding platinum agents to the same standard NACT regimen of weekly paclitaxel (with or without carboplatin) followed by anthracycline plus cyclophosphamide.^[18,19,26] For the analysis, only patients from the CALGB study who were included in the 2 arms without bevacizumab were considered. In these 3 RCTs, 234 out of 418 patients (46.0%) achieved a pCR following NACT. Among them, 169 out of 308 patients (56.0%) in the platinum-based chemotherapy group, and 146 out of 408 patients (35.8%) in the platinum-free chemotherapy group, achieved a pCR (OR 2.27, 95% CI 1.72–3.00, P < .001; I² = 66.0%, P = .005) (Fig. 3C). Table S4, Supplemental Digital Content, https://links.lww.com/MD/Q134, available at Annals of Oncology online, details the sensitivity analysis.

3.2. Objective response rate

Six RCTs reported the ORR at the conclusion of NACT.^{[18,21,23,24,26,28]} The evaluation methods employed in these RCTs are detailed in Table S5, Supplemental Digital Content, https://links.lww.com/MD/Q134, accessible online in the Oncology Yearbook. Overall, the ORR following NACT was reported as 82.8% (1212 of 1464 patients). Specifically, in the platinum-based chemotherapy group, the ORR was 88.4% (648 of 733 patients), whereas in the platinum-free chemotherapy group, the ORR was 77.2% (564 of 731 patients). The OR for these findings was calculated at 2.31 with a 95% CI ranging from 1.74 to 3.09 (P < .001; I² = 66%, P = .01) (Figure S1, Supplemental Digital Content, https://links.lww.com/MD/Q135). The sensitivity analysis is available in Table S6, Supplemental Digital Content, https://links.lww.com/MD/Q134, which can be accessed online in the Annals of Oncology.

3.3. EFS, OS, DDFS, iDFS, SD, and PD

Several RCTs have reported survival outcomes. Some RCTs exclusively reported OS, while others focused solely on EFS, DDFS, or iDFS. Specifically, 4 RCTs documented EFS outcomes.^{[19,21,24,29]} The median follow-up duration was 27 months (range: 5–61 months) in the ChiCTR-TRC-1900023776 study,^[29] 37 months (range: 11–52 months) in the NeoCART study,^[24] 39 months in the CALGB 40603 study, and 47.5 months in the GeparSixto; GBG 66 study. Eight RCTs reported OS data.^{[18–22,24,27,28]} The median follow-up was 47.7 months in the NACATRINE study,^[22] 48.2 months (range: 31.1–60.0 months) in the ChiCTR-TRC-14005019 study,^[28] 47.0 months (range: 1.6–61.5 months) in the GeparOcto-dGBG 84 study,^[20] 60 months in the WSG-ADAPT-TN study,^[27] and 54 months in the BrighTNess study.^[18] Additionally, 3 RCTs reported DDFS outcomes,^[20,25,27] with a median follow-up duration of 58.2 months in the RSF-2017-00000557 study.^[25] Two RCTs reported iDFS outcomes.^[20,27]

In the NeoCART study,^[24] among the patients in the DCb group, 10 achieved CR, 32 achieved PR, 1 developed SD, and 1 experienced PD. In the EC-D group, 9 patients achieved CR, 31 patients achieved PR, 4 developed SD, and no cases of PD were reported. In the ChiCTR-TRC-14005019 study,^[28] the TEL group saw 44 patients achieving CR, 48 achieving PR, and 7 developing SD, with no cases of PD. In contrast, in the TE group, 31 patients achieved CR, 44 patients achieved PR, 23 developed SD, and 3 patients experienced PD.

In summary, platinum-based chemotherapy improved EFS (hazard ratio [HR] 0.68, 95% CI 0.53–0.88, P = .03; I² = 0.0%, P = .507) (Fig. 4A) compared to non platinum-containing chemotherapy, however, no significant differences were observed in OS (HR 0.84, 95% CI 0.68–1.05, P = .119; I² = 15.3%, P = .310) (Fig. 4B), DDFS (HR 0.97, 95% CI 0.71–1.31, P = .839; I² = 28.2%, P = .248) (Fig. 4C), and iDFS (HR 0.88, 95% CI 0.65–1.19, P = .396; I² = 22.5%, P = .256) (Fig. 4D).

Figure 4.

Hazard ratios for event-free survival (A), overall survival (B), distant disease-free survival (C), and invasive disease-free survival (D) of platinum-based (Platinum) versus platinum-free (Controls) NACT. The size of the squares is proportional to the weight of each study. CI = confidence intervals, HR = hazard ratio, NACT = neoadjuvant chemotherapy.

3.4. Grade 3 and 4 AEs

In this study, we identified a total of 14 AEs; however, it is important to note that these events did not necessarily occur concurrently across all trials. The OR was 2.41 (95% CI 2.11–2.74, P < .001; I² = 85%, P < .001). Figure 5 provides an overview of the safety profile for grade 3 and 4 AEs comparing the platinum-based chemotherapy group to the platinum-free chemotherapy group.

Figure 5.

Safety profile overview. Odds ratios for grade 3 and 4 neutropenia, grade 3 and 4 anemia, grade 3 and 4 thrombocytopenia, and grade 3 and 4 neuropathy in the platinum-based (Platinum) versus platinum-free (Controls) NACT. CI = confidence interval, NACT = neoadjuvant chemotherapy, OR = odds ratio.

The GeparSixto; GBG66, UMIN000003355, and GeparOcto-dGBG 84 RCTs did not report specific rates of AEs in the subgroup of TNBC patients.^[20,21,26] Nevertheless, the results of these 3 RCTs were included in this safety analysis, as the risk of AEs is not expected to be influenced by hormone receptor status. Figure 6 displays an overview of the safety profile for grade 3 and 4 AEs in the comparison between platinum-based and platinum-free chemotherapy groups.

Figure 6.

Safety profile overview. Odds ratios for grade 3 and 4 anemia, grade 3 and 4 leukopenia, grade 3 and 4 neuropathy, grade 3 and 4 neutropenia, and grade 3 and 4 thrombocytopenia in the platinum-based (Platinum) versus platinum-free (Controls) NACT. CI = confidence interval, NACT = neoadjuvant chemotherapy, OR = odds ratio.

3.5. Neutropenia

Nine RCTs reported the rates of grade 3 and 4 neutropenia.^{[18,19,22–25,27–29]} Overall, 530 out of 1861 patients (28.5%) developed grade 3 and 4 neutropenia following neoadjuvant treatment. Specifically, 358 of 920 patients (38.9%) in the platinum-based chemotherapy group experienced neutropenia, compared to 172 of 941 patients (18.3%) in the platinum-free chemotherapy group (OR 3.00, 95% CI 2.41–3.74, P < .001; I² = 88%, P < .001) (Figure S2, Supplemental Digital Content, https://links.lww.com/MD/Q135 and Table S7, Supplemental Digital Content, https://links.lww.com/MD/Q134, available at Annals of Oncology online).

3.6. Thrombocytopenia

Similarly, 9 RCTs reported the rates of grade 3 and 4 thrombocytopenia.^{[18,19,22–25,27–29]} A total of 116 out of 1907 patients (6.1%) developed grade 3 and 4 thrombocytopenia after neoadjuvant treatment. Within this cohort, 103 of 900 patients (11.4%) in the platinum-based chemotherapy group experienced thrombocytopenia, while only 13 of 917 patients (1.4%) in the platinum-free chemotherapy group did so (OR 8.31, 95% CI 4.78–14.44, P < .001; I² = 36%, P = .16) (Figure S3, Supplemental Digital Content, https://links.lww.com/MD/Q135 and Table S8, Supplemental Digital Content, https://links.lww.com/MD/Q134, available at Annals of Oncology online).

3.7. Anemia

Seven RCTs reported the incidence of grade 3 and 4 anemia.^{[18,22,24,25,27–29]} Overall, 109 out of 1336 patients (8.2%) developed grade 3 and 4 anemia following neoadjuvant treatment. In the platinum-based chemotherapy cohort, 94 of 655 patients (14.4%) experienced this adverse effect, compared to 15 of 681 patients (2.2%) in the platinum-free chemotherapy group. The OR was calculated to be 9.05 (95% CI 5.04–16.24, P < .001), with a low heterogeneity (I² = 3%, P = .39) (Figure S4, Supplemental Digital Content, https://links.lww.com/MD/Q135 and Table S9, Supplemental Digital Content, https://links.lww.com/MD/Q134, available online at Annals of Oncology).

3.8. Leukopenia

Six RCTs reported the rates of grade 3 and 4 leukopenia.^{[18,19,25,27–29]} In total, 168 of 1535 patients (10.9%) developed grade 3 and 4 leukopenia after neoadjuvant treatment. Among those receiving platinum-based chemotherapy, 100 of 759 patients (13.2%) were affected, whereas 68 of 776 patients (8.8%) in the platinum-free chemotherapy group experienced this condition. The OR for leukopenia was 1.56 (95% CI 1.11–2.19, P = .01), with moderate heterogeneity (I² = 18%, P = .30) (Figure S5, Supplemental Digital Content, https://links.lww.com/MD/Q135 and Table S10, Supplemental Digital Content, https://links.lww.com/MD/Q134).

3.9. Neuropathy

Five RCTs reported the incidence of grade 3 and 4 neuropathy.^{[18,19,22,23,27]} Overall, 31 out of 1312 patients (2.4%) developed grade 3 and 4 neuropathy following neoadjuvant treatment. In the platinum-based chemotherapy group, 18 of 647 patients (2.9%) experienced this adverse effect, while 13 of 665 patients (2.0%) in the platinum-free chemotherapy group were affected. The OR for neuropathy was calculated to be 1.36 (95% CI 0.67–2.73, P = .39), with moderate heterogeneity (I² = 38%, P = .19) (Figure S6, Supplemental Digital Content, https://links.lww.com/MD/Q135 and Table S11, Supplemental Digital Content, https://links.lww.com/MD/Q134, available online at Annals of Oncology).

4. Discussion

This meta-analysis comprehensively summarizes existing evidence from RCTs evaluating the activity, efficacy, and safety of platinum-based chemotherapy as a NACT for TNBC patients. We observed that platinum-based NACT is associated with significantly higher pCR and EFS rates but at the cost of greater blood toxicity risk, while no significant differences were observed in OS, DDFS, and iDFS.

Platinum demonstrates preferential efficacy against low-aggressiveness tumor clones (thereby increasing pCR rates), but exhibits limited impact on highly resistant lesions (which ultimately determine survival outcomes). The persistence of residual disease following intensive regimens (e.g., platinum-nab-paclitaxel followed by anthracyclines) may indicate intrinsic chemoresistant biology, making these patients ideal candidates for novel therapeutic approaches. This hypothesis aligns with observations from WSG-ADAPT-TN: non-pCR patients in the carboplatin arm showed poorer prognosis; suggests treatment efficacy may enrich for more aggressive biological subtypes under therapeutic pressure. However, these findings remain hypothetical and require further verification. Additionally, despite achieving pCR, some patients still experience recurrence, while many who do not attain pCR remain disease-free. These observations may explain the discordance between higher pCR rates and unimproved OS.

Compared with other neoadjuvant therapies, platinum combined with anthracycline- and taxane-based chemotherapy was more effective neoadjuvant treatment for TNBC in terms of pCR improvement. Platinum salts serve as DNA-damaging agents that demonstrate enhanced efficacy in tumors characterized by defective DNA repair mechanisms. The platinum salts interact with DNA within cells, causing distortions in the DNA double helix and inducing both single-strand breaks and double-strand breaks. When these damages cannot be efficiently repaired, it results in cell death.^[30] These agents have demonstrated efficacy in cancers harboring germline BRCA mutations, given the crucial role that BRCA1/2 proteins play in repairing DNA damage.^[31–33] A significant proportion of TNBC displays a BRCA-like phenotype (BRCAness), suggesting that these tumors exhibit a high sensitivity to platinum salts.^[34,35] Encouraging findings from clinical studies and the present network meta-analysis corroborate the neoadjuvant use of platinum-containing, anthracycline- and taxane-based chemotherapy in TNBC.

Regarding the role of platinum agents, between-treatment estimations showed that pCR benefits from platinum-containing chemotherapy were generally more pronounced than those of the platinum-free counterpart. The present meta-analysis, including 12 RCTs (N = 2650), provides updated evidence on the debated role of platinum-based NACT in TNBC patients. A significant absolute 16.7% increased pCR rate was observed with the use of platinum-based NACT. This is consistent with previous meta-analysis that the addition of platinum agents to neoadjuvant therapies further improved pCR in TNBC.^[36]

Despite being limited by the relatively small number of studies involved, the results from the indirect analysis of time-to-event endpoints demonstrated the strongest correlation between platinum-based chemotherapy combined with anthracycline and taxane and an extended EFS. Combined with the findings from the latest meta-analysis which showed platinum-based NACT significantly increased EFS as compared with platinum-free regimens.^[37] However, whether the combination is associated with improved OS, DDFS, and iDFS remains to be seen.

The practice-altering success of NACT utilizing platinum in combination with anthracycline and taxane was accompanied by a notable rise in premature treatment discontinuation, primarily attributed to AEs related to the treatment. In NACATRINE, there was no difference in the rate of permanent treatment discontinuation between groups (34.2% vs 32.1%, P = .488). In GeparOcto-dGBG 84, 2 deaths under therapy occurred in the PM(Cb) arm due to pneumonia and multiple septic cerebral embolism. In NCT01276769, grade 3 and 4 thrombocytopenia was present in 4 patients (8.5%) in the platinum arm, resulting in 2 cases of treatment termination. In UMIN000003355, in the PCb-CEF arm, 26 patients discontinued CP due to AEs and 3 patients discontinued CEF due to AEs. Five and 6 patients in the P-CEF arm discontinued P and CEF, respectively, due to AEs. In GeparSixto; GBG 66, the interruption rate in the carboplatin group was 49% compared to 36% in the non-carboplatin group (P = .023). In our meta-analysis, we collected data on a total of 14 AEs, among which neutropenia, thrombocytopenia, anemia and leukopenia were found to be more prevalent. The utilization of platinum-based chemotherapy varies widely, and as of the time of this article’s composition, the National Comprehensive Cancer Network guidelines do not routinely advocate for its employment. The absence of a demonstrable long-term survival benefit serves as the rationale for the National Comprehensive Cancer Network’s restriction on the routine application of platinum-based chemotherapy. Though there are certainly increased hematological toxicities associated with platinum chemotherapy, grade 3/4 neuropathy was not different between groups.^[38] For patients intolerant to AEs, de-escalation of the chemotherapy backbone may be considered. Hematological toxicity is one of the most common and potentially life-threatening side effects in clinical practice. Its impact on clinical decision-making is multifaceted and dynamic, extending throughout the entire treatment process. During the treatment cycle, if blood counts fail to recover to safe thresholds by the scheduled chemotherapy day, treatment must be postponed; prolonged delays may compromise efficacy. Furthermore, when severe (grade 3/4) adverse reactions occur, chemotherapy drug doses typically need to be reduced in the subsequent cycle. Severe fatigue, anemia, recurrent infections, bleeding risks, and resulting hospitalizations significantly diminish patients’ quality of life and may lead to treatment discontinuation or patient refusal to continue therapy. Additionally, repeated hospitalizations, blood transfusions, supportive medications (such as thrombopoietin receptor agonists), extra testing, and monitoring substantially increase healthcare costs.

A few sub-analyses show high heterogeneity, and we will now analyze the following reasons. First, although all RCTs selected platinum-based chemotherapy agents, the specific types of platinum drugs varied across studies (some used lobaplatin, while others chose carboplatin). This may be one of the reasons for the high heterogeneity. Second, while platinum-based chemotherapy was evaluated in all studies, the dosages and treatment schedules were not entirely consistent. Third, some RCTs had relatively small sample sizes, which could amplify the observed heterogeneity. In our study protocol, we prespecified the method for assessing heterogeneity (I² statistic) and the handling strategy (subgroup analysis). We calculated the I² value, which clearly indicated substantial heterogeneity. Additionally, we conducted sensitivity analyses by sequentially excluding each study to observe changes in the pooled effect size, CIs, and heterogeneity levels. Finally, despite the substantial heterogeneity observed, sensitivity and subgroup analyses demonstrated that the core outcomes and their statistical significance remained relatively stable. Therefore, while the observed heterogeneity warrants cautious interpretation of the results, we conclude that it does not fundamentally alter the key findings of this study.

The presence of network meta-analysis had several limitations. Firstly, uncertainty surrounded all estimates due to the heterogeneity among eligible studies in terms of patient populations, treatment durations, and drug dosages. Consequently, rigorous inclusion criteria were implemented for the selection of eligible studies. Sensitivity analyses were also performed to guarantee the robustness and reliability of indirect inferences. Nevertheless, TNBC remains a notably heterogeneous disease, necessitating further characterization of the target patient population for the clinical application of our findings. Secondly, certain comparisons were based on a limited number of patients, leading to effect sizes that were constrained by broad 95% CIs and posing a potential risk of introducing publication bias. Thirdly, while we fully acknowledge the prognostic significance of BRCA status in breast cancer, only 3 of the included RCTs reported BRCA status data. Consequently, a systematic analysis could not be performed in this study. We look forward to incorporating more BRCA-status-reported RCTs in future updates of this research.

5. Conclusions

In conclusion, our meta-analysis showed that platinum-based NACT was associated with significantly increased pCR and EFS rates in TNBC patients. Although it entails an increased risk of grade 3/4 hematological toxicity, the risk is still manageable. These findings suggest that the addition of a platinum agent to standard neoadjuvant anthracycline- and taxane-based chemotherapy may be considered an option in TNBC patients.

Author contributions

Conceptualization: Xiaobo Zhao.

Data curation: Jiangzhuo Wu, Xiao Yan, Jiang Fang.

Formal analysis: Jiangzhuo Wu.

Funding acquisition: Lin Peng.

Methodology: Lin Peng, Xiaobo Zhao.

Project administration: Lin Peng.

Resources: Lin Peng.

Software: Jiangzhuo Wu.

Supervision: Xiao Yan, Jiang Fang, Xiaobo Zhao.

Validation: Xiaobo Zhao.

Visualization: Jiang Fang.

Writing – original draft: Jiangzhuo Wu.

Supplementary Material

medi-104-e44740-s001.pdf ^{(164KB, pdf)}

medi-104-e44740-s002.pdf ^{(974.1KB, pdf)}

medi-104-e44740-s003.pdf ^{(567.4KB, pdf)}