Survival outcomes of neoadjuvant versus adjuvant chemotherapy in triple-negative breast cancer: a meta-analysis in 36480 cases.

Background The survival outcomes of neoadjuvant chemotherapy(NACT) versus adjuvant chemotherapy(ACT) in triple-negative breast cancer (TNBC) remains unclear. Therefore, in this study, a meta-analysis was conducted to analyze current evidence on the survival outcomes of neoadjuvant versus ACT in TNBC. Methods A systematic search of PubMed and Embase databases were done to identify relevant articles investigating the survival outcomes of NACT versus ACT in TNBC. Results A total of nine studies involving 36480 patients met the selection criteria.Of them,10728(29.41%) received NACT, and 25752(70.59%) received ACT.The pathological complete response (pCR) rate is 35% (95% CI = 0.23-0.48). Compared with ACT, the overall survival(OS) of NACT was poor(HR = 1.59; 95%CI = 1.25-2.02; P=0.0001), and there was no significant difference in disease-free survival(DFS) between them(HR = 0.85; 95%CI = 0.54-1.34; P=0.49). NACT with pCR can significantly improve the OS(HR = 0.53; 95%CI = 0.29-0.98; P=0.04) and DFS(HR = 0.52; 95%CI = 0.29-0.94; P=0.03), while the OS(HR = 1.18; 95%CI = 1.09-1.28; P<0.0001) and DFS(HR = 2.36; 95%CI = 1.42-3.89; P=0.0008) of patients with residual disease(RD) following NACT were worse compared with ACT. Conclusion The findings suggest that, for TNBC, Patients receiving NACT with pCR can significantly improve OS and DFS compared with ACT and it turns to be opposite when patients receiving NACT with RD.


Introduction
Breast cancer is the most common cancer in women. Globally, there are 1.2 million to 1.4 million women with breast cancer and about 400,000 patients die of breast cancer [1]. TNBC is defined as a type of breast cancer lacking the three most common types of receptors namely, estrogen receptor(ER), progesterone receptor(PR) and HER-2 expression which are known to fuel most of the breast cancer growth. TNBC accounts for 12%-20% of all breast cancers and it is characterized by high pathological grade, strong invasiveness, local recurrence, high metastasis rate, and poor prognosis [2][3][4]. Therefore, it is very important to give systemic treatment as soon as possible. In the 3 past, ACT has been the standard treatment for TNBC, but now more and more patients with TNBC have adopted NACT because NACT can control systemic micrometastases, reduce the tumor burden, provide surgical or conservative breast surgery opportunities for locally advanced breast cancer patients, and detect tumor sensitivity to chemotherapeutic drugs. pCR after NACT is associated with an improved tumor-free survival rate of patients [5,6] and Studies have shown that the rate of NACT pCR for TNBC is significantly higher than that of non-TNBC [7][8][9][10][11]. This also seems to indicate that TNBC is more suitable for NACT. At present, studies have compared the prognosis of NACT and ACT in patients with TNBC, but the results are contradictory [12][13][14]. So whether TNBC can get better survival outcomes from NACT than ACT is still controversial. Our study was to compare the survival outcomes of NACT versus ACT in TNBC by meta-analysis.

Material And Methods Search strategy
A systematic literature search of PubMed and Embase databases for the period up to January 18, 2020, identified eligible studies. The keywords used in the search strategy were triple-negative breast neoplasms OR triple-negative breast cancer OR triple-negative breast carcinoma AND neoadjuvant OR preoperative AND Adjuvant chemotherapy OR chemotherapy. A total of nine articles met the eligibility criteria with a total of 36480 patients [12][13][14][15][16][17][18][19][20]. The identified studies were not limited to any geographical or linguistic barriers. Because of the small number of included studies, publication bias was not assessed.

Inclusion and Exclusion Criteria
Eligible studies had to meet the following inclusion criteria in order to ensure that only high-quality studies were considered for this analysis. Inclusion criteria: 1) patients diagnosed with TNBC; 2) the study compared the survival outcomes of NACT with ACT; 3) the study must assess the overall prognosis of TNBC; 4) the study reporting survival outcomes in terms of OS and/or DFS. The exclusion criteria were as follows: 1) articles lacking the original data; 2) studies lacking information on survival outcomes in TNBC; 3) articles not reporting or giving an estimate of the hazard ratio (HR) and a 95% confidence interval (95% CI). Figure 1 illustrates the eligibility criteria of articles used for this research. 4 Data Extraction and statistical analysis A standardized data extraction form was used to extract relevant information from each of the studies. For each of the eligible studies, the following information was extracted: The first author's name, publication year, patients' countries, study design, number of participations, tumor stage and follow-up results.
Meta-analysis was conducted using RevMan version 5.3 (RevMan, version 5.3 for Windows; Cochrane Collaboration, Oxford, UK). The hazard ratios (HRs) with 95% CIs were calculated to estimate the association between the DFS and OS of NACT and ACT in TNBC. Published data and Kaplan-Meier survival curves were used to extract the HR estimates according to the methods reported by Tierney et al if the HRs were not directly provided [21]. Chi-squared based Q-test (P > 0.10) and I 2 were used to determine statistical heterogeneity within the studies. When I 2 < 50%, the studies were considered to have acceptable heterogeneity and the fixed-effects model was used. Otherwise, the randomeffects model was used. All P values were two-sided, and P < 0.05 was considered to be statistically significant. Figure 1 shows the process of selecting the included studies. A total of 420 articles were first identified for evaluation. 158 were excluded because they were duplicates while 238 were irrelevant to this study. However, based on the inclusion and exclusion criteria described, 15 were excluded. Therefore, a total of 9 publications were eligible for the meta-analysis. Out of the 9 studies selected, 7

Results
were retrospective while 2 were prospective. Table 1 presents the characteristics of the included studies. Table 1 Main characteristics and results of the eligible studies.
The DFS after NACT versus ACT in TNBC was evaluated in the Clifton, Sharma and Philipovskiy trials 6 [12,15,20]. A total of 569 patients were included in the pooled analysis with a median follow-up time of 5.14 years. There is no statistically difference between the NACT and ACT arms on the DFS( HR = Same as our study,most studies have confirmed that patients with TNBC have a better prognosis after achieving pCR in NACT [7,28].Although patients who received NACT may have more serious diseases, we still found that NACT with pCR significantly improved survival. This suggest that in our study, compared with all patients receiving NACT, the survival advantage of ACT is determined by the part of patients with residual disease after NACT. In our meta-analysis, the pCR rate is 35% (95% CI = 0.23-0.48; P < 0.01) (Fig. 2). In these studies, patients with early stage, small tumor and negative lymph node are more likely to obtain pCR. Only the study by Clifton showed a pCR rate of 54% with all the rest of the studies having pCR rate below 50%. Therefore, the high rate of RD is associated with a poor survival rate of NACT.
In our study, there is no statistically significant difference in the DFS between the ACT and ACT arms.
It's different to other study. A study involving 4,756 breast cancer patients showed that women who received NACT had higher local recurrence rates within 15 years (21.4% vs 15.9%) than women who received ACT, and the risks were significantly different ( RR 1.37; 95% CI = 1.17-1.61;P = 0.0001) [29].
Among them, the risk increased significantly after 0-4 years (RR 1.35; 95% CI = 1.11-1.64) and 5-9 years (RR = 1.53; 95% CI = 1.08-2.17). Women who received NACT in their study were more likely to take breast-conserving treatments than women who received ACT (65% vs 49%) .The higher rate of breast-conserving surgery after NACT may lead to a significantly higher risk of local recurrence. Mauri validated this by conducted a meta-analysis of 9 randomized trials involving 3,946 patients, the results show that the local recurrence risk of the NACT group is significantly higher than that of ACT group due to the higher breast-conserving surgery rate in NACT(RR = 1. 22; 95% CI = 1. 04 − 1. 43; P = 0. 018) [30]. This may be related to the disunity of tumor regression model after NACT, the difficulty of tumor localization, and the increased difficulty of breast-conserving surgery [31,32].But in our study, there was no significant difference between NACT and ACT in breast conserving surgery rate 8 (RR = 0.84; 95% CI = 0.57-1.23; P = 0.37) (Fig. 3).This can be used to explain why our study has different results from other studies.
This meta-analysis has some limitations. One of the limitations is that 8 of the 9 studies we included analyzed the overall survival benefits of TNM stage I-III patients without distinguishing the early and late stages of the disease, so our analysis could not compare the survival benefits according to different stages of the disease.As proved in this study, TNBC with negative lymph node, small tumor and early stage is more likely to get pCR in NACT. The poor survival benefit of NACT compared with ACT is determined by patients with RD. If all patients were restricted to the early stage of disease, we speculated that it might be possible to obtain a higher pCR rate and a better survival benefit for NACT compared to ACT. Despite this limitation, our study also concluded that NACT with pCR can significantly improved survival in TNBC. It gives us some hints and thoughts, such as how to screen out the patients with TNBC who may achieve pCR with NACT to improve survival. In addition, The HR and 95% CI extracted from the survival curves may be less reliable than those directly obtained from the articles.

Conclusion
The meta-analysis showed that the OS of NACT is worse than that of ACT in TNBC, and there is no significant difference in DFS between them. However, when patients achieving pCR with NACT, OS and DFS can be significantly improved.Our results suggest that the selection of NACT treatment may be more effective in patients predicted to achieve pCR, while ACT is for patients who cannot achieve pCR and this can improve the patient's survival rate. However, more trials of the prognosis of NACT versus ACT in TNBC are still needed to confirm this result. Not applicable.

-Consent to publish
Not applicable.

-Availability of data and materials
All the data are available without restriction. Researchers can obtain data from the corresponding author.

-Competing interests
The author(s) declare that they have no competing interests.

-Funding
Not applicable.