Introduction

Triple-negative breast cancer (TNBC), defined by the absence of immunostaining for estrogen and progesterone receptors and lack of overexpression or amplification of human epidermal growth factor receptor 2, accounts for approximately 15–20% of all breast cancers1,2. Compared with hormone receptor-positive breast cancers, TNBC has a worse prognosis, with most events occurring relatively early. More than 50% of patients experience relapse in the first 3–5 years after diagnosis3. Studies have demonstrated that TNBC is more responsive to systemic chemotherapy than are other molecular subtypes, and taxane-containing and anthracycline-containing chemotherapy is currently the main systemic treatment option for this type of breast cancer4,5. Adjuvant chemotherapy (AC) and neoadjuvant chemotherapy (NAC) yield similar disease-free survival and overall survival (OS) rates6,7,8.

Multiagent chemotherapy regimens have shown benefit in terms of improving both disease-specific survival and OS in patients with early TNBC whether administered before or after surgery and are widely recommended to reduce the risk of relapse9,10. The international guidelines list sequential use of a taxane and anthracycline plus cyclophosphamide as one of the preferred regimens for early TNBC11. However, despite treatment with anthracycline-taxane-based therapy and adjuvant capecitabine in patients with residual disease at surgery, 5-year metastasis-free survival remains at approximately 70%, with approximately 30–40% of patients developing metastatic disease and dying of their cancer12,13,14. Therefore, improvements in long-term treatment outcomes are needed for patients with early-stage TNBC.

NAC also provides an opportunity to collect biological and prognostic information during the response to guide further treatment. Although AC remains more common globally, NAC is now considered standard practice and its use is becoming more frequent15. NAC is the current standard of care for patients with early disease10,13,16,17. The short-term goal of NAC in patients with TNBC is a pathological complete response because of its association with prolonged event-free survival and OS7,18. The agents in the chemotherapy regimens do not differ between NAC and AC other than use of pembrolizumab in the neoadjuvant setting for high-risk patients. However, NAC allows pathological response-guided adjuvant treatment that can improve survival. Furthermore, NAC provides the opportunity to downstage the primary tumor and axillary nodes, allows for de-escalation of locoregional interventions, and the tumor response can help guide the need for subsequent AC and foster drug discovery19. Studies have shown that NAC with atezolizumab in combination with nab-paclitaxel and anthracycline-based chemotherapy significantly improves the pathological complete response rate20. However, in a series of 194 patients with TNBC, similar locoregional and distant recurrence-free survival rates were found for those with T1mic/T1a or T1b tumors (94.5% vs 95.5%), with no differences in outcomes according to whether or not chemotherapy was administered (95.9% vs 94.5%)21. Patients with extensive residual invasive cancer after NAC are at high risk of recurrent metastatic disease but whether survival is better in patients with small T1 TNBC who receive NAC than in those who receive AC after upfront surgery is uncertain. NAC is the standard for TNBC higher than T1 stage while data on the efficacy of NAC in T1 TNBC are unclear. The primary objective of this study was to compare OS and BCSS in patients with clinical T1 TNBC who have undergone upfront surgery followed by AC with that in those who have undergone NAC followed by surgery.

Results

Baseline characteristics

Application of the inclusion criteria identified 6249 eligible women with cT1 TNBC between 2010 and 2015, including 5404 (86.5%) who underwent surgery followed by AC (the AC group) and 845 (13.5%) who received NAC followed by surgery (the NAC group). The patient selection process is shown in Fig. 1. The disease characteristics at baseline are compared between the NAC group and the AC group in Table 1. Significant differences in the distributions of age, T stage, N stage, and surgery were observed between the study groups. Most of the patients were aged over 50 years, of white ethnicity, and married, had grade 3 T1c N0 stage tumors, and had undergone breast-conserving surgery (BCS). More than half of the patients had received radiotherapy. Compared with patients in the AC group, those in the NAC group tended to be younger than 50 years and had a higher lymph node-positive rate. After PSM to balance confounding factors, the χ2 test suggested balance and comparability between the two groups.

Figure 1
figure 1

Flowchart of patient selection.

Full size image
Table 1 Baseline characteristics of patients with cT1 triple-negative breast cancer.
Full size table

Overall survival outcomes

Kaplan–Meier curves were generated for the NAC and AC groups. Before and after PSM (Fig. 2A,C), OS was poorer in the NAC group. Univariate Cox analyses revealed significant between-group differences in age, marital status, disease grade and stage, type of surgery, and type of adjuvant therapy received. In view of the potential for bias stemming from mutual associations between these variables, the subsequent Cox proportional hazards regression analysis was performed, and the results are shown in Table 2. The risk of death was significantly higher in the NAC group than in the AC group (HR 1.46, 95% CI 1.21–1.76; p < 0.001). OS was poorer in patients who were aged 50 years or older than in patients who were younger (HR 1.27, 95% CI 1.07–1.50; p = 0.006) and poorer in patients who were single, divorced, separated, or widowed than in patients who were married (HR 1.28, 95% CI 1.10–1.49; p = 0.002). As expected, a higher tumor grade was associated with a worse prognosis. The HR was 2.17 (95% CI 1.28–3.69; p = 0.004) for T1c disease in comparison with T1a disease; however, there was no significant difference between T1b and T1a disease (HR 1.47, 95% CI 0.84–2.57; p = 0.179). The prognosis was significantly worse in patients with positive lymph nodes than in those with negative lymph nodes (HR 4.79, 95% CI 3.81–6.00; p < 0.001). Furthermore, OS was worse in patients who underwent radical mastectomy than in those who underwent BCS (HR 1.27, 95% CI 1.06–1.52; p = 0.008). However, there was no significant survival benefit from radiotherapy.

Figure 2
figure 2

Comparison of OS and BCSS between for patients with cT1 triple-negative breast cancer receiving NAC and upfront surgery + AC. (A) Before PSM, overall survival; (B) Before PSM, breast cancer-specifc survival; (C) After PSM, overall survival; (D) After PSM, breast cancer-specifc survival.

Full size image
Table 2 Prognostic factors associated overall survival using Cox proportional hazards analysis for patients with cT1 triple-negative breast cancer.
Full size table

Breast cancer-specific survival outcomes

The Kaplan–Meier curves show that the prognosis of BCSS with better in the AC group than in the NAC group (Fig. 2B,D). Table 3, which summarizes the results of the multivariable analysis, shows that BCSS was significantly poorer in the NAC group than in the AC group (HR 1.54, 95% CI 1.26–1.89; p < 0.001), which was consistent with the findings for OS. However, unlike with the OS results, age and marital status was not identified as an independent risk factor for BCSS, and no significant difference in BCSS was found between the grade and surgery subgroups. BCSS was poorer in patients with T1c BC than in those with T1a BC (HR 1.93, 95% CI 1.09–3.43; p = 0.029) but no significant difference was found between patients with T1b and those with T1a breast cancer (HR 1.34, 95% CI 0.73–2.46; p = 0.349). As expected, node-positive disease was associated with worse BCSS (HR 6.40, 95% CI 5.01–8.18; p < 0.001). Furthermore, radiotherapy was not associated with improved BCSS.

Table 3 Prognostic factors associated breast cancer-specific survival using Cox proportional hazards analysis for patients with cT1 triple-negative breast cancer.
Full size table

Subgroup analysis stratified by age, T stage, and N stage

To further investigate the effect of NAC on survival, the population was divided into T1a, T1b, T1c, N0, N1 and N2-3 subgroups according to T stage and N stage, and baseline statistics and PSM were performed for each subgroup (Supplementary Tables 1–6). In different T1 stages, both before and after PSM, the OS and BCSS of the NAC group were worse than those of the AC group (Fig. 3A–L). In the T1N0 group, there were no statistically significant differences in OS and BCSS between the AC and NAC groups before and after PSM (Fig. 4A,D,G,J). However, in the lymph node-positive subgroup, both before and after PSM, the OS and BCSS of the NAC group were inferior to those of the AC group (Fig. 4B,C,E,F,H,I,K,L).

Figure 3
figure 3

Comparison of OS for patients with cT1 triple-negative breast cancer between receiving NAC and upfront surgery + AC based on clinical T category. (A) Before PSM, OS of T1a; (B) Before PSM, OS of T1b; (C) Before PSM, OS of T1c; (D) Before PSM, BCSS of T1a; (E) Before PSM, BCSS of T1b; (F) Before PSM, BCSS of T1c; (G) After PSM, OS of T1a; (H) After PSM, OS of T1b; (I) After PSM, OS of T1c; (J) After PSM, BCSS of T1a; (K) After PSM, BCSS of T1b; (L) After PSM, BCSS of T1c.

Full size image
Figure 4
figure 4

Comparison of OS for patients with cT1 triple-negative breast cancer between receiving NAC and upfront surgery AC based on clinical N category. (A) Before PSM, OS of N0; (B) Before PSM, OS of N1; (C) Before PSM, OS of N2-3; (D) Before PSM, BCSS of N0, (E) Before PSM, BCSS of N1; (F) Before PSM, BCSS of N2-3; (G) After PSM, OS of N0; (H) After PSM, OS of N1; (I) After PSM, OS of N2-3; (J) After PSM, BCSS of N0, (K) After PSM, BCSS of N1; (L) After PSM, BCSS of N2-3.

Full size image

We stratified all cases according to age, T stage, and N stage and performed multivariate analyses to evaluate the prognostic association of NAC in selected demographic and pathological subgroups (Table 4). Receiving NAC was associated with worse OS and BCSS across age groups. Similarly, NAC was associated with worse OS and BCSS across the various T stages than in patients who received AC. NAC was also associated with poorer OS and BCSS across the N stages in patients with positive lymph nodes. In patients with negative lymph nodes, there was no significant difference in OS or BCSS between the NAC group and the AC group.

Table 4 The adjusted hazard ratio for OS and BCSS associated with NAC and AC in patients with different ages and cancer stages.
Full size table

Discussion

In this retrospective analysis of data for patients with T1 TNBC identified in a large nationwide US database, 86.5% (n = 5404) received AC and 13.5% (n = 845) received NAC. OS and BCSS were worse in the patients who received NAC than in those who received AC. We compared the prognostic impact of the two treatments in various demographic and clinical subgroups and found that, for patients with triple-negative breast cancer at T1N0 stage, there was no significant difference in OS and BCSS between NAC and AC. However, NAC was associated with worse OS and BCSS in patients with T1 TNBC who were clinically node-positive. This suggests that even in cases of lymph node positivity, the indications for neoadjuvant chemotherapy should be carefully considered for TNBC patients with T ≤ 2 cm.

AC as the traditional treatment of T1 TNBC, the main purpose is to eradicate residual microscopic cancer cells after surgery, further reducing the risk of recurrence and metastasis22. AC can significantly improve survival rates in patients with T1 TNBC and lymph node involvement or other strong risk factors23. Moreover, evidence is now emerging from clinical trials suggesting that surgery followed by chemotherapy might not be the optimal strategy to maximize the chances of survival in patients with TNBC. Therefore, it is necessary to explore the effectiveness of NAC. The core goal of NAC is to effectively reduce the tumor volume preoperatively, thereby downstaging the tumor, increasing eligibility for BCS, and decreasing the need for axillary dissection. Furthermore, NAC can effectively remove small metastases, thereby significantly reducing the risk of postoperative recurrence and metastasis and creating more favorable conditions for patient rehabilitation and follow-up treatment. Patients presenting with node-negative tumors measuring < 2 cm do not usually require downstaging because most are already candidates for BCS22,23. However, T1 TNBC, although relatively small, is usually aggressive and prone to recurrence, and NAC may be a suitable option if the tumor is large and lymph node metastasis or other risk factors are present. Consistent with these observations, 79.1% of the women with T1 TNBC who received NAC in our study had T1c disease and 41.3% were lymph node-positive.

NAC provides an opportunity to directly observe the efficacy of a particular chemotherapy regimen, however there are currently no biomarkers to reliably predict response to systemic therapy for TNBC, and patients who have extensive residual invasive cancer after NAC are at high risk of recurrent metastatic disease. TNBCs often achieve pCR by neoadjuvant chemotherapy. With modern anthracycline and taxane containing regimens, pCR rates are reported in 30–40% of the cases24,25. Several strategies have been pursued to increase the pCR rate in TNBC. The addition of platinoids increases the probability of pCR by about 45%26. This moderate increase is associated with more complexity and side effects and, thus far, has not yet resulted in an advantage in long-term outcomes25,27,28. Therefore, there is some controversy regarding the choice of NAC or AC for T1 TNBC, and the choice may need to be individualized after a detailed assessment of the patient. Our data argues that although NAC allows for a tailored approach to systemic therapy after breast surgery, providing more potentially toxic neoadjuvant systemic therapy may not be in the patient’s best interest in small TNBC. Compared with NAC, patients in the AC group were associated with better OS and BCSS, while avoiding the risk of PD from NAC, and adjuvant therapy was tailored to accurate pathological staging.

However, it is important to note that NAC and AC are not mutually exclusive. In some cases, patients may need both NAC and AC to maximize the effectiveness of treatment, AC may be necessary for patients with lymph node metastasis or other risk factors identified in the postoperative pathology report, and patients failing to achieve pCR after a full course of neoadjuvant chemotherapy can have their prognosis improved with additional treatments9. The NAC group in our study included patients who received chemotherapy both before and after surgery. Comparison of OS and BCSS in patients with stage T1 TNBC according to whether they received NAC or AC showed that in most patients, NAC was associated with poorer survival, suggesting that NAC may not be suitable for patients with T1 TNBC.

Through univariate and multivariate Cox regression analyses, we identified age, tumor size, lymph node as robustly associated with survival outcomes. The size of the tumor is an important prognostic indicator for triple-negative breast cancer, as it directly affects the degree of lymph node involvement and the risk of recurrence and metastasis. Axillary lymph node metastasis also significantly impacts the patient's prognosis23,29,30. Our data showed no benefit from neoadjuvant chemotherapy in T1N0 TNBC, while in node-positive T1 TNBC, neoadjuvant chemotherapy increased the risk of death compared to adjuvant chemotherapy. We further analyzed the difference in survival between the two treatment modalities in tumors of different sizes, and the results showed that even in patients with T1c, neoadjuvant chemotherapy significantly increased the risk of death. In addition, neoadjuvant chemotherapy was associated with poorer OS and BCSS in either age group. In summary, for T1 TNBC, systemic responses and side effects should be weighed even if the tumor is larger than 10 mm or combined with lymph node metastasis, especially in the absence of other risk factors, NAC should be carefully considered.

Limitations

This study has several limitations. First, despite our efforts to adjust for multiple confounders, the inherent bias of retrospective studies is difficult to eliminate by existing analytic instruments. Second, the variables used to determine the treatment sequence did not include information on all treatments that were administered before or after surgery, only that “systemic therapy” was given. Systemic therapy is not limited to chemotherapy and does not indicate the specific chemotherapy regimens provided. Third, there is a time delay in SEER data, which may not reflect the latest treatment methods. Finally, the clinical characteristics and therapeutic interventions used were limited to the US population and might be different in other countries.

Methods

Data source and study population

The study data were obtained from the SEER database. The SEER program is sponsored by the National Cancer Institute, contains data collected from 18 population-based cancer registries between 1975 and 2016, and provides deidentified cancer statistics for the US population31. This study was deemed to be exempt from Institutional Board Review at the participating research institutions.

Data for 15,195 patients identified to have T1 TNBC between 2010 and 2015 were extracted for evaluation of the association of NAC with survival. Patient staging was confirmed as per the seventh edition of the American Joint Committee on Cancer (AJCC) staging system (2010). Tumors with a largest diameter of ≤ 2 cm in the gross pathological specimen were defined as stage T1 and classified as T1a (1–5 mm), T1b (6–10 mm), or T1c (11–20 mm). Data for patients with T1mic and Tis were excluded. We selected patients with one primary malignant neoplasm only and retrieved their demographic and clinical data, including age, sex, race/ethnicity, marital status, tumor grade, Seventh AJCC T stage, surgery, chemotherapy, and radiotherapy. Patients with incomplete information were excluded, as were male patients.

Statistical analysis

Frequencies and proportions were calculated to describe the baseline characteristics of eligible patients, and categorical variables were compared between the different stages using Pearson’s chi-squared test. OS and BCSS were investigated in the total study population. To improve the comparability between groups, 1:1 propensity score matching (PSM, method = "nearest", caliper = 0.02) was used to reduce the influence of confounding factors. Kaplan–Meier survival was generated and log-rank tests applied to determine the influence of NAC on the prognosis. In further univariate and multivariable Cox proportional analyses, hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) were calculated to evaluate the association of NAC with OS and BCSS. Next, to verify the National Comprehensive Cancer Network guidelines for TNBC, we calculated adjusted HRs and corresponding 95% CIs for patients who received NAC in comparison with those who received AC when stratified by age and T and N stage to handle potential biases and evaluate the prognostic association with NAC. All statistical analyses and generation of figures were performed using R version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). A two-sided p-value of < 0.05 was considered statistically significant.

Ethics declarations

The SEER database is publicly accessible and does not provide personal identifying information. Therefore, institutional review board approval and informed patient consent were not required.