Postoperative Radiotherapy Contributes to the Survival Benefit of Breast-Conserving Therapy over Mastectomy

Purpose A survival benefit of breast-conserving therapy (BCT) over mastectomy has been shown in recent studies. This study aimed to explore differences in recurrence patterns between BCT and mastectomy and clarify the contribution of radiotherapy (RT) to the survival benefit of BCT. Methods Consecutive patients with pT1-2/pN0-1/M0 breast cancer between 2009 and 2015 in our institution were retrospectively reviewed and compared in matched cohorts using 1 : 1 propensity score matching (PSM). Results A total of 2370 patients were enrolled with a median follow-up of 75 (3–148) months. In the cohort without regional nodal irradiation (RNI), WBI was associated with significantly increased 10-year relapse-free survival (RFS), distant metastasis-free survival (DMFS), and regional recurrence-free survival (RRFS) compared with mastectomy alone. There were 419 pairs in the cohort without RNI and 87 pairs in the cohort with RNI after PSM. In the PSM cohort, improved 10-year RFS (95.4% vs. 82.7%, p < 0.05), DMFS (97.4% vs. 84.1%, p < 0.05), and RRFS (99.1% vs. 95.5%, p < 0.05) were observed in WBI compared with mastectomy alone. Regarding the first recurrence event, WBI demonstrated a significantly lower cumulative rate of distant metastases than mastectomy alone. There was no significant difference in survival outcomes between WBI plus RNI and PMRT before and after the PSM. In patients without RNI, mastectomy alone was significantly associated with unfavorable RFS (HR = 2.3, 95% CI 1.2–4.5, p < 0.05) and DMFS (HR = 2.5, 95% CI 1.1–5.8, p < 0.05). Conclusion This study found the benefit of RFS and DMFS in BCT patients compared with those treated with mastectomy without RNI but not in those treated with RNI. We hypothesized that RT played an important role in reducing the risk of regional recurrence and distant metastases.

Recent population-based studies of contemporary treatment have found superior survival with BCT over mastectomy in early breast cancer [5][6][7][8][9][10][11][12]. Van Maaren and colleagues found better 10-year distant metastasis-free survival (DMFS) of BCT over mastectomy-alone in T1N0 patients and most patients in this subgroup received no adjuvant systemic therapy [6]. Onitilo et al. reported a better survival of BCT over mastectomy-alone and hypothesized an association of RT with therapeutic benefts [13]. One explanation for this fnding is that the application of the modern RT technique spares normal tissue [14]. Although the 15-year follow-up of the EORTC 22922 study confrmed the advantage of regional nodal irradiation (RNI) in breast cancer recurrence and mortality risk, its role was not evaluated in randomized studies comparing BCT with mastectomy [15]. Locoregional recurrence (LRR) and distant metastasis (DM), as two important surrogates proved to be reduced by 50% with RT in a recently published EBCTCG meta-analysis, were not reported in previous studies [16]. Tese fndings imply that modern RT technique might play an increasing role in the survival beneft of BCT over mastectomy.
Currently, although BCT remains the frst treatment recommendation for T1-2, N0-1 BC patients in international guidelines, its therapeutic outcomes in comparison with mastectomy and the efect of RT on survival beneft remain to be explored and this is the aim of the current study.

Study Population.
Consecutive patients with pT1-2, pN0-1, and M0 breast cancer who were treated with BCT or mastectomy between 2009 and 2015 in Ruijin Hospital were retrospectively reviewed. BCT consisted of BCS and whole breast irradiation (WBI). Patients with bilateral breast cancer, histologically confrmed sarcoma, receiving neoadjuvant systemic therapy, without WBI after BCS or without RNI in PMRT were excluded. Tis study was approved by the institutional Medicine Review Board, and a waiver of consent was obtained due to its retrospective nature.
Patient-related, tumor-related, and treatment-related characteristics were collected. Te tumor was staged according to the American Joint Committee on Cancer (AJCC) Cancer staging manual (7th edition). Estrogen receptor (ER) and progesterone receptor (PR) statuses were assessed using immunohistochemical (IHC) analysis and considered positive if ≥10% of cells staining positive. A positive human epidermal growth factor receptor 2 (HER2) status was defned as an expression level intensity of 3+ on IHC or a gene amplifcation ratio >2.2 by fuorescence in situ hybridization (FISH). ER, PR, HER2, and Ki-67 were used to classify patients into 5 molecular subtypes: luminal A-like, luminal B-like (HER2-negative), luminal B-like (HER2positive), HER2-positive (nonluminal), and triple-negative breast cancer (TNBC) according to the St Gallen 2013 consensus guidelines [17].

Treatment.
Te surgical treatment options (BCS or mastectomy) were based on the tumor characteristics and patient's preference. Sentinel lymph node biopsy (SLNB) was performed in patients with clinically negative lymph nodes (LNs), and axillary lymph node dissection (ALND) was performed in patients with clinically positive LNs or positive SLNB.
Te choice of adjuvant therapy was decided by a multidisciplinary team consisting of breast surgeons, medical oncologists, radiation oncologists, and others when needed as previously described [18].
Te dose prescription to the whole breast/chest wall with or without regional nodes was 45-56 Gy in 25-28 fractions. A tumor bed boost (TBB) was delivered with a dose of 10-16 Gy in 5-8 fractions following WBI. Te supra/infraclavicular CTVs are kept constant for RNI in patients with positive lymph nodes, internal mammary nodes and low axilla at the discretion of the radiation oncologist.

Statistical Analysis.
Local recurrence (LR) was defned as any recurrence in the ipsilateral breast or chest wall. Regional recurrence (RR) was defned as recurrence in the ipsilateral axillary, supra/infraclavicular, or internal mammary nodes. Distant metastasis (DM) was defned as any evidence of metastatic disease beyond the locoregional region. Relapse-free survival (RFS) events included LR, RR, DM, and death from breast cancer. All statistical endpoints were calculated from the date of surgery to the defned events, including breast cancer-specifc survival (BCSS), local recurrence-free survival (LRFS), regional recurrencefree survival (RRFS), distant metastasis-free survival (DMFS), and RFS.
Categorical variables were assessed using Pearson's χ 2 test or Fisher's exact test as appropriate. One-to-one propensity score matching (PSM) analysis was performed to control for selection bias between the 2 cohorts: a cohort without RNI (WBI versus mastectomy-alone) and a cohort with RNI (WBI plus RNI versus PMRT). Te patients were matched on age, pT/pN stage, pathology, grade, molecular subtypes, adjuvant chemotherapy, endocrine therapy, and targeted therapy. Survival curves were estimated using the Kaplan-Meier method and compared by the log-rank test. A Cox proportional hazards model was used for multivariate analysis, including predictive variables. All hazard ratios and hazard ratio confdence interval bounds for the categorical variables were rounded to a single decimal place. A twosided p value <0.05 was considered statistically signifcant. Statistical tests were performed using SPSS version 22.0.

Patient Characteristics.
A total of 2370 patients were enrolled, of whom 520 (21.9%) received BCT and 1850 (78.1%) received mastectomy ± RT. Among patients treated with BCT, 429 (82.5%) received WBI, and 91 (17.5%) received WBI plus RNI. Among patients treated with mastectomy, 220 out of 1850 (11.9%) patients received PMRT. Te patient characteristics are shown in Table 1. Compared with those treated with mastectomy, patients who received BCT were more likely to be <50 years of age and have pT1N0 or pT1N1 disease. Te distribution of molecular subtypes also varied signifcantly between WBI and mastectomy alone. In addition, the proportion of patients with hormone receptor (HR) positive who received endocrine therapy was higher in WBI than in mastectomy.

Terapeutic Outcomes in the Cohort with RNI (WBI plus RNI vs. PMRT).
In the cohort with RNI, no signifcant diference between WBI plus RNI and PMRT in terms of BCSS, RFS, DMFS, RRFS, and LRFS was found. After PSM, 87 pairs of patients were included (Table 1). Tere was no diference in BCSS, RFS, DMFS, RRFS, or LRFS rates at 10 years in the matched cohort ( Figure 2). Similarly, the patterns of the frst recurrence between WBI plus RNI and PMRT were comparable (Table 3).

Discussion
In the present study, we compared the survival outcomes of early-stage breast cancer patients who received BCS with WBI or mastectomy according to all parameters related to survival. We initially divided the patients into four groups according to the treatment strategy (BCS + WBI, mastectomy-alone, BCS + WBI with RNI, and mastectomy + PMRT), and our study found that improved survival was observed only in the comparison of BCT vs. mastectomy alone. Prior to the present study, several researchers attributed the survival beneft of BCT over mastectomy to postoperative radiotherapy after BCS without providing information on the feld of radiation or the recurrence rates [11,13,19,20]. Onitilo and colleagues made comparisons between BCS and mastectomy and found that OS was similar between BCS and mastectomy regardless of RT, but OS was better in the BCTgroup than in the mastectomy alone group, which was likely to be related to the addition of RT [13]. A study from the Netherlands found a survival beneft in the T1N0 subgroup in which most patients did not receive adjuvant systemic therapy after BCT or mastectomy-alone; therefore, the improvement in survival in the BCT group was attributed to RT [6]. Our study further investigated whether RT might play an essential role in the survival beneft from risk reduction of RR and DM in BCT. Kaplan-Meier survival analysis demonstrated that there was a signifcantly   Figure 1: Comparisons of breast cancer-specifc survival, relapse-free survival, distant metastasis-free survival, regional recurrence-free survival, and local recurrence-free survival curves between WBI and mastectomy alone after PSM. possibly had no infuence on survival. It could be hypothesized that RTcontributed greatly to the risk reduction of RR and DM, resulting in the beneft of RRFS and DMFS in BCT over mastectomy.
Our results are inconsistent with previous outcomes of randomized trials conducted in the 1980s showing equivalent survival between BCT and mastectomy [1][2][3][4]. Te inconsistency may be explained by the fact that these trials were conducted before the modern era, and screening programs, surgical techniques, systemic therapies, and RT techniques have developed dramatically. A pooled analysis found that mastectomy signifcantly reduced the risk of locoregional recurrence when compared with BCT in previous trials [21], but the rate of local recurrence after BCT has fallen over time [22]. In addition, all of the BCT patients in our analysis received TBB after whole breast irradiation, which would reduce the risk of local recurrence [23], while the BCT patients only received postoperative radiotherapy without TBB in the NSABP B-06 trial [2].
Te trend toward a decreased risk of regional recurrence associated with RT between WBI and mastectomy alone is probably due to the lower incidence of axillary recurrence as frst recurrence, which was halved in the WBI group compared with mastectomy alone (0.7% vs. 1.4%) in the PSM cohort. Tis is in line with other studies providing evidence of decreased axillary recurrence in BCS plus RT compared with mastectomy without RT, even in low-risk patients [19,20]. In addition, the results from the ACOSOG Z0011 and IBCSG-23-01 indicated that the omission of ALND did not compromise 10-year OS in 1-2 SLNB-positive or SLNB micrometastasis patients [24,25], which refected potential protection against axillary recurrence by incidental irradiation originating from WBI. It was estimated that more than 50% of axillary level I and 20%-30% of level II receives 95% of the prescribed radiation dose using tangential felds [26].
Given that subclinical radiation dose to the axilla contributes signifcantly to the regional control, a decreased risk of distant metastasis in BCT with WBI patients over mastectomy alone probably accounts for the major part of the survival beneft, as our results showed better DMFS of WBI over mastectomy only, which had been found in previous studies [20,27,28]. In both the MA.20 and EORTC 22922 studies, a decreased rate of DM with the addition of RNI was found [29,30]. Tis was possibly attributed to a direct link between the decreased risk of regional recurrence and possible subsequent distant metastasis. Te results from the EBCTCG meta-analysis showed that RT after BCS reduces the risk of recurrence and death from breast cancer, indicating that RT plays an important role in killing microscopic tumor foci after surgery and therefore reduces the potential for both local recurrence and distant metastasis [16]. Overgaard et al. found that the 10-year disease-free survival was 48% among women with PMRT and 34% among those without PMRT in high-risk patients (p < 0.001) and a similar magnitude of beneft in 10-year OS from 45% without PMRT to 54% with PMRT (p < 0.001), showing that the addition of RT reduces locoregional recurrences and subsequent distant metastasis that improves survival [31]. However, although we found better DMFS in WBI compared with mastectomy alone, no improvement in BCSS was found. Te fact that this did not translate into better BCSS might be explained by the salvage treatment after distant metastases and the 75 months of median follow-up time, which was not long enough to detect diferences in the specifc survival. T1-2 breast cancer patients with negative LNs account for a large proportion of breast conservation. As RT is not routinely recommended in this group of patients when undergoing mastectomy, patients receiving BCS are more likely to be provided the risk reduction of RR and DM with subsequent RT. Improved survival associated with BCT compared with mastectomy alone was also found in the early stage, node-negative breast cancer in another study [10]. For patients with 1-3 positive LNs, expert consensus on the role of PMRT remains controversial [32]. A meta-analysis from EBCTCG found that PMRT reduced locoregional recurrence and mortality rates in this group [33], whereas some retrospective analyses observed no signifcant diferences in BCSS or overall survival (OS) with the addition of PMRT in the era of modern systemic therapy [34,35]. Due to the debate on this issue, a small portion of patients with pN1 undergoing mastectomy did not receive PMRT in our study and might have received RT if BCS had been used. Although other studies have also shown improved survival with BCT compared with mastectomy in node-positive patients [36,37], our study is the frst to separately compare BCS + WBI-only vs. mastectomy without PMRT and BCS + WBI with RNI vs. mastectomy with PMRT including RNI. By comparing these two cohorts with PSM analysis and multivariate analysis, in early-stage patients, mostly T1-2N0, BCS + WBI showed a protective therapeutic efect in reducing the risk of regional recurrence and distant metastasis, which should be attributed to the RT. In high-risk patients who received comprehensive local-regional RT, regardless of the pattern of primary breast surgery, no signifcant diferences in terms of local-regional recurrence or distant metastasis were found. One can hypothesize that BCS with WBI is associated with better survival than mastectomy alone; therefore, BCS is strongly recommended in early breast cancer if indicated.   Figure 2: Comparisons of breast cancer-specifc survival, relapse-free survival, distant metastasis-free survival, regional recurrence-free survival, and local recurrence-free survival curves between WBI plus RNI and PMRT after PSM.
As a retrospective study, selection bias was inevitable. PSM analysis balanced the major prognostic parameters, but some variables remained unbalanced. For example, the role of the oncotype DX recurrence score (RS) was not assessed in our study, although its role in guiding BCS is still uncertain [38]. Te median follow-up in our study was 75 months and longer follow-up is necessary. Finally, whether the role of RT in survival varies in diferent molecular subtypes, as previously reported, was not explored in our study [37,[39][40][41].

Conclusions
Our study confrms that in early-stage breast cancer not indicated for comprehensive RNI, BCS with WBI is associated with a better prognosis than mastectomy without PMRT in terms of RFS, DMFS, and RRFS. For high-risk patients with comprehensive local-regional RT, similar survival outcomes are observed regardless of the type of breast surgery. Our results, along with the previous studies fnding a survival advantage with BCT, provide evidence that this efect should be attributed to RT since it reduces the risk of regional recurrence and distant metastases rather than to the type of surgery itself. Our fndings support BCT as a therapeutic choice of priority in early breast cancer.

Data Availability
Te data are not available due to ethical restrictions.

Disclosure
Te funding sources had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Conflicts of Interest
Te authors declare that they have no conficts of interest.