Breast-conserving surgery versus mastectomy for treatment of breast cancer after neoadjuvant chemotherapy

Background To compare recurrence and survival outcomes between breast-conserving surgery (BCS) and mastectomy after neoadjuvant chemotherapy (NACT). Methods The data of 730 patients who underwent NACT between 2000 and 2014 were retrospectively reviewed. A total of 104 (14.2%) patients received BCS and 626 (85.8%) received mastectomy. Locoregional recurrence (LRR), distant metastases (DM), disease-free survival (DFS), breast cancer–specific survival (BCSS), and overall survival (OS) were analyzed using the Kaplan–Meier method. The impact of BCS versus mastectomy on outcomes was assessed by multivariate Cox models. Inverse probability of treatment weighting (IPTW) was used to balance covariates between the two groups. Results The median follow-up of BCS and mastectomy groups were 86.5 and 87.4 months, respectively. There were significant differences in distribution of most baseline characteristics between two groups. Compared with those who underwent mastectomy, the patients with BCS had similar 5-year LRR, DM, and DFS rates, but had significantly higher 5-year BCSS (98.9% vs. 90.4%, P = 0.005) and OS (98.9% vs. 90.1%, P = 0.003) rates. Multivariate analysis also showed that BCS significantly improved BCSS (HR = 0.27, 95% CI: 0.08-0.85, P = 0.025) and OS (HR = 0.25, 95% CI: 0.08-0.79, P = 0.018). After IPTW adjustment, the LRR, DM, DFS, BCSS and OS between two groups had no significant differences. Conclusions The recurrence and survival outcomes are comparable with BCS and mastectomy. Thus, BCS is a safe treatment option for selected breast cancer patients after NACT.


Introduction
Several prospective randomized clinical trials have demonstrated that breast-conserving surgery (BCS) plus radiotherapy can provide long-term survival outcomes comparable to that with mastectomy (1-3) and better cosmetic outcomes and quality of life (4-6) in early-stage breast cancer (BC). Some recent large population-based studies have even shown better survival rates with BCS (7-10).
Traditionally, neoadjuvant chemotherapy (NACT) was a standard treatment of unresectable locally advanced breast cancer (LABC) to convert these patients to candidates for surgery. Two prospective randomized trials, the NSABP B18 and EORTC 10902, compared NACT and adjuvant chemotherapy in operable BC, and reported similar survival outcomes between two groups (11,12). Moreover, these two trials both also found that NACT increased the rate of BCS by reducing the size of primary tumor. Based on the findings, NACT has been more widely accepted for operable BC and to facilitate BCS for the patients with large tumors who are initially considered for mastectomy (13,14). In addition, it has potential to reduce resected volumes for the patients who are already candidates for BCS and achieve better cosmetic outcomes (15). Meanwhile, NACT can determine the chemo-sensitivity and reduce micrometastasis.
However, it is a main concern that whether BCS after NACT would increase the rate of ipsilateral breast tumor recurrence. A meta-analysis including ten studies (from 1983 to 2002) by the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) found that 69% of patients achieved complete or partial clinical response after NACT and that the frequency of BCS increased from 49% to 65% after NACT. Nevertheless, the 15-year local recurrence rate was higher with NACT than with adjuvant chemotherapy (21.4% vs. 15.9%) (16). Other studies also showed unexpectedly high local recurrence rates for patients who received BCS after NACT (17,18). One of the most important challenges for surgeons is to determine the original tumor location and excision extent to obtain tumor-free margins and achieve good cosmetic outcomes when performing BCS after NACT, especially for the patients with good response to NACT but whose residual tumor cells are scattered over the residual volume of disease. A consensus has been reached among the experts that "no ink on tumor" guideline was an adequate resection margin for BCS after NACT, because no relationship has been found between the margin width and outcomes (19,20). Nowadays, the safety of BCS after NACT for the operable LABC remains controversial. Therefore, we conducted this study to compare the oncological outcomes after BCS versus mastectomy in patients receiving NACT.

Study population
The data of patients treated with NACT between 2000 and 2014 in our institution were retrospectively reviewed. The inclusion criteria were 1) receiving NACT with large primary tumor and/or heavy axillary lymph nodal burden; 2) breast cancer with cT1-3N0-2M0 stage and ypT0-2N0-2M0 stage; 3) information available on estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) status; and 4) information available on whether or not to receive adjuvant treatment, such as chemotherapy, radiotherapy, endocrine therapy, or HER2-targeted therapy. The exclusion criteria were 1) received BCS without adjuvant radiotherapy; 2) mastectomy patients with proven ypN1-2 stage but no postmastectomy radiotherapy (PMRT); 3) relapse within 2 months; or 4) failure to complete at least 6 months of follow-up after surgery. We also excluded the patients with cT4/ypT3-4/ypN3 disease, because none of them had received BCS in our initial cohort. This study was approved and the need for informed consent was waived by Ethics Committee of National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (approval number: 15-057/984), as this was a retrospective analysis of chart data.

Outcomes
Locoregional recurrence (LRR) was defined as recurrence in the ipsilateral breast or chest wall, ipsilateral axilla, supra-or infraclavicular lymph nodes, or internal mammary lymph nodes. Distant metastasis (DM) was defined as evidence of metastatic disease beyond the locoregional regions. Disease-free survival (DFS) was calculated from the date of definitive surgery to the date of LRR, DM, death, or the last follow up. Breast cancer-specific survival (BCSS) was calculated from the date of start of NACT to the date of death from BC or last follow-up. Overall survival (OS) was calculated from the date of start of NACT to the date of death or the last follow up.

Statistical analysis
The chi-square test was used to compare patient characteristics between the BCS and mastectomy groups. Multivariate Cox regression models were used to assess the impact of surgery methods on recurrence and survival after adjusting for confounding factors including treatment era, age, clinical stage, NACT cycles, response to NACT, histological grade, lymphovascular invasion (LVI), molecular subtype/trastuzumab, ypStage (pathologic stage after NACT), hormone receptor (HR) status/endocrine therapy, and adjuvant chemotherapy. To reduce the effect of selection bias and potential confounding factors, the differences in baseline covariates between BCS and mastectomy groups were balanced by the inverse probability of treatment weighting (IPTW) method (21). The IPTW approach attempts to mimic a situation in which treatment is randomly allocated to individuals. The Kaplan-Meier method was used for analysis of recurrence and survival before and after IPTW; the log-rank test was used for comparisons between the groups. Statistical analysis was performed using the SPSS 22.0 (IBM Corp., Armonk, NY, USA) and R 4.1.2 (https://www.r-project.org/). Two-sided P <.05 was considered statistically significant.
PMRT was recommended for patients with ypN1-2 stage and for ypN0 patients with high-risk factors (i.e., age < 45 years, cT3, cN2, presence of LVI, or ER/PR negative status

Treatment outcomes
The median follow-up of BCS and mastectomy groups were 86.5 months (range, 6.5-169.9 months) and 87.4 months (range, 6.1-201.3 months), respectively. A total of 58 (7.9%) patients had developed LRR, 159 patients (21.8%) had developed DM, and 83 (11.4%) patients had died. Among the 83 patients who died, 78 (94.0%) died of the BC and 5 (6.0%) died of other causes. All five patients that died from other causes were in the mastectomy group; the causes of death included pulmonary fibrosis (n = 1), acute pancreatitis (n = 1), leukemia (n = 1), anemia and thrombocytopenia (n = 1), and unknown cause (n = 1). The patient who died from acute pancreatitis had hypertension, gastric ulcer, and hyperthyroidism; the other four patients had no comorbidity at the time of diagnosis of BC.

Discussion
In this single-center cohort study, we retrospectively evaluated the oncologic safety of BCS compared with mastectomy following NACT in patients with BC and found that LRR, DM, DFS, BCSS, and OS are comparable with BCS and mastectomy. Overall, our findings suggest that BCS is a safe and effective treatment option after NACT for patients with BC.
In recent years, NACT has become standard treatment for unresectable or resectable LABC and is being increasingly used in early-stage BC. The NSABP B18 trial examining the sequencing of chemotherapy demonstrated that patients treated with NACT had similar survival outcomes with those treated with adjuvant chemotherapy and NACT could increase BCS rates by tumor down-staging (11). The ipsilateral breast tumor recurrence (IBTR) rates were comparable in the NACT and adjuvant chemotherapy groups (7.9% vs. 5.8%). However, among women being converted from mastectomy to BCS candidates after NACT, the IBTR rate was higher (14.3%). Thus, the effectiveness of BCS after NACT remains unclear. Chen et al. reported acceptably low 5-year IBTR and LRR rates of 5% and 9% for the patients treated with BCS after NACT in a cohort of 340 patients from the MD Anderson Cancer Center; however, women with any one of the high-risk factors (i.e., cN2-3, pathologically residual tumor > 2 cm, multifocal residual disease, and LVI) had higher IBTR and LRR rates (9%-13% and 16%-23%) (18).
In our study, the surgical method was selected based on tumor response to NACT and patients' preference. Given the small breast size for most eastern women and few applications of oncoplastic surgery, the guidelines recommended that BCS should be offered to the patients who had tumor size < 3cm or had appropriate ratio of tumor to breast volume to achieve good cosmesis. In our study, 621 (85.1%) patients had tumor size ≥ 3cm who were not good candidates for BCS initially. After NACT, 563 (77.1%) had pathologic tumor size < 3cm who were presumed to be appropriate for BCS. However, most patients selected to receive mastectomy, which constituted a comparable cohort for the present study. In the present study, the LRR, DM, and DFS were similar in BCS and mastectomy patients; BCSS and OS were better in BCS patients, although the differences were not statistically significant after confounding factors were adjusted by IPTW. A meta-analysis found no significant difference in local and regional recurrence between BCS and mastectomy patients after NACT and that BCS patients had lower incidence of DM and better DFS and OS (22). Previous studies that reported the oncologic safety of BCS after NACT are summarized in Table 3 (13,(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36). All studies were retrospective and most had small sample size. In a cohort of 561 patients treated with NACT, Simons et al. found significantly better DFS and OS after BCS than after mastectomy, but the statistical significance disappeared after correcting for confounders (36). An analysis of population-based data from the New Jersey State Cancer Registry (NJSCR) demonstrated that BCS patients after NACT had significantly better 10-year BCSS than mastectomy patients, and the difference remained even in propensity-matched comparison. However, the study findings must be interpreted cautiously because all patients did not receive radiotherapy after mastectomy and some important clinical factors such as tumor size, radiotherapy and chemotherapy data were missing in the database (35). Barranger et al. reported a 72.3% "mastectomy to BCS" conversion rate after NACT for resectable LABC. For whom mastectomy was the only conceivable surgical option initially, the 5year DFS and OS of BCS and mastectomy patients were comparable (74% vs. 59% and 77% vs. 77%, respectively). But baseline tumor characteristics were not balanced between the two groups, the BCS patients having smaller tumor size and higher pCR rate, which were not adjusted during survival analysis and would have affected the outcomes (13). Most of the aforementioned studies reported results that were consistent with ours, i.e., that BCS after NACT is a safe alternative to mastectomy in patients with BC; however, these   earlier studies had obvious limitations such as unbalanced baseline characteristics and missing important data. In our study, we included patients who had received adjuvant radiotherapy per the current guideline (e.g., radiotherapy after BCS, PMRT if ypN1-2 or ypN0 with high-risk factors). Confounding due to differences in baseline covariates between the two surgical groups was minimized by using the IPTW method. These measures make our results more convincing. All five patients who died from non-breast-cancer causes were in the mastectomy group, but most had no comorbidities at diagnosis of BC; therefore, the poorer OS in the mastectomy group might not be related to presence of comorbidities. In multivariate analysis, ypStage was independent prognostic factors for most oncologic outcomes, and hormone receptornegative status was a significant predictor of poor OS. As other authors have also demonstrated that advanced post-NACT stage and triple-negative status were significant predictors of poor outcome; this is not surprising since these factors indicate aggressive disease (30,37).
Though this study suggests that BCS is a safe treatment option for patients after NACT, some concerns remain. In clinical practice, the primary tumor site is usually difficult to locate after tumor regression. One study that assessed pathological response of BC to NACT found increased incidence of multifocality and in situ lesions localized within the original tumor-bearing area after tumor shrinkage (38); this can lead to difficulty in defining the extent of resection necessary to achieve safe margin during BCS. However, over the past few years, there have been major advances to improve the probability of safe BCS after NACT. These advances include increased application of breast MRI, use of metal markers to improve definition of tumor location, and improved detection of multifocal or multicentric tumor, greater attention to achieving pathologically negative BCS margins and use of modern radiotherapy techniques that provide more precise dose coverage and thus improve local control and decrease toxicity. A policy review endorsed by several European societies and clinical trial groups has provided a practical working toolbox for the surgical treatment of early-stage BC after NACT (39). There is now consensus that all patients receiving NACT must undergo comprehensive evaluation in multidisciplinary team meetings, undergo imaging by multiple modalities (e.g., MRI and ultrasound) at diagnosis, and have clips placed at the primary site before NACT. In addition, response assessment at different time points must be done by the same imaging modality used at initial diagnosis; careful preoperative evaluation for localization, volume excision, and retrieval of breast markers is essential before BCS. Precise margin assessment and appropriate radiotherapy are also important for successful BCS. Through the close cooperation of multidisciplinary team, not only do the patients with resectable LABC have an opportunity to be converted from mastectomy to BCS candidates after NACT, but the women with early-stage triple-negative or HER2 -positive BC who are currently candidates for NACT are safe to receive BCS.
Some limitation of this study should be acknowledged. First, the retrospective design might have introduced a selection bias. Although IPTW was used to balance known variables in the two groups, it is possible that other confounders were unevenly distributed. Second, the 15-year span of patient inclusion was very long; but the patients in different treatment era between the two groups were comparable, and the influences of the changes in the diagnosis and treatment of BC over this period between the two groups were similar, e.g. the proportion of the patients who had HER2 -positive disease but did not received trastuzumab-targeted therapy was comparable after IPTW between the two groups. Third, the findings of this study can only be applied to specific populations, i.e. patients with cT1-3N0-2M0 and ypT0-2N0-2M0 BC, and the effect of BCS in patients with more advanced stage remains to be accessed. To our knowledge, this is the largest study to compare the outcomes between BCS and mastectomy after NACT, thus, we believe that our study makes a meaningful contribution to clinical practice.

Conclusions
Breast-conserving surgery appears to be a safe treatment option for selected BC patients after neoadjuvant chemotherapy. It does not compromise locoregional, distant control, DFS, BCSS and OS compared with mastectomy.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement
The studies involving human participants were reviewed and approved by Ethics Committee of National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (approval number: 15-057/984). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.
Author contributions Y-CS: Formal analysis, investigation, data collection, methodology, and writing of the first draft. ZH: statistical analysis, investigation, data collection, methodology. HF, YuT, HJ, Y-WS, JJ, Y-PL, BC, YuanT, S-NQ, N-NL, and NL: Patient care and review, and editing of the manuscript. Y-XL and S-LW: Study design, formal analysis, validation, statistical analysis guidance, patient care, and editing of the manuscript. All authors contributed to the article and approved the submitted version.

Funding
This research was funded by the CAMS Innovation Fund for Medical Sciences (CIFMS) (2021-I2M-1-014), the