Prognostic Factors and Surgery for Breast Cancer Patients With Locoregional Recurrence: An Analysis of 5,202 Consecutive Patients

Purpose With the application of “less extensive surgery” in breast cancer treatment, the pattern of locoregional recurrence (LRR) has significantly changed. This study aims to evaluate the risk and prognostic factors of LRR in a recent large breast cancer cohort. Methods Consecutive early breast cancer patients who received surgery from January 2009 to March 2018 in Shanghai Ruijin Hospital were retrospectively analyzed. LRR was defined as recurrence at the ipsilateral breast (IBTR), chest wall, or regional lymph nodes and without concurrent distant metastasis (DM). Patients’ characteristics and survival were compared among these groups. Results Among 5,202 patients included, 87 (1.7%) and 265 (5.1%) experienced LRR and DM as first event after a median 47.0 (3.0–122.5) months’ follow-up. LRR was significantly associated with large tumor size and positive lymph node status (p < 0.05). Forty (46.0%) patients received further salvage surgery after LRR and had a significantly better 3-year post-recurrence overall survival than those who did not (94.7% vs. 60.7%, p = 0.012). Multivariate analysis showed that salvage surgery for LRR was independently associated with better survival (HR = 0.12, 95% CI 0.02–0.93, p = 0.043) along with estrogen receptor (ER) positivity (HR = 0.33, 95% CI 0.12–0.91, p = 0.033). Conclusion LRR rate was relatively low in recent era of breast cancer treatment. Tumor size and lymph node status were associated with risk of LRR, and salvage surgery for selected LRR patients achieved an excellent outcome.


INTRODUCTION
Breast cancer is the most commonly diagnosed cancer and the leading cause of cancer mortality in females worldwide (1). With a better understanding of tumor biologic behavior, innovations in screening techniques, and the development of comprehensive multidisciplinary treatment strategies, more breast cancers can be diagnosed at early stages. Less extensive surgery, for instance, breast-conserving surgery (BCS) followed by radiotherapy and sentinel lymph node biopsy (SLNB) in selected patients demonstrated equivalence with mastectomy and axillary lymph node dissection (ALND) in terms of survival but with less comorbidities (2,3).
Locoregional recurrence (LRR) is a clinically relevant, predominant pattern of treatment failure in breast cancer. LRR patterns vary across initial surgical approach and mainly involve recurrence in chest wall post-mastectomy, residual breast after BCS, or regional lymph nodes (LNs). According to previous evidence, factors associated with increased risk of LRR include young age at diagnosis, greater tumor size, involvement of regional LN, high histological grade, triple negative (TN) or human epidermal growth factor receptor 2 (HER2)-positive subtype, lack of endocrine therapy, and omitting indicated adjuvant radiotherapy (4)(5)(6). Different from the palliative management of distant metastasis (DM), salvage surgery plays an important role in the comprehensive management of LRR. Patients who received salvage surgery for LRR reported relatively satisfactory 5-year overall survival (OS) ranging from 40.8% to 90.9% (7,8), suggesting that selective LRR patients would benefit from salvage surgery and quite a number of LRR patients could be cured. However, retrospective series showed that between 15% and 37% patients with LRR had concurrent DM at the time of presentation (9)(10)(11)(12)(13)(14)(15)(16)(17)(18). Disease outcomes and treatment strategies of these populations can be very different from those with LRR alone. The management of LRR should be based on systemic evaluation and be discussed in a multidisciplinary setting.
However, studies of LRR were mostly conducted in the late 1990s to early 2000s and in western populations. Following the change of initial surgical procedures from "maximal tolerable treatment" to "minimal effective treatment," the pattern of LRR has also significantly changed. With an increasing rate of BCS and SLNB, now we meet more patients with ipsilateral breast tumor recurrence (IBTR) and regional LN recurrence in clinical practice. To this end, the objective of this study was to analyze the risk and prognostic factors of LRR in the current "less extensive surgery" era.

Patients
We retrospectively included consecutive female patients diagnosed with primary invasive breast cancer and received radical surgical treatment from January 2009 to March 2018 in Comprehensive Breast Health Center, Shanghai Ruijin Hospital. Patients with complete clinicopathological information, with at least 3 months of follow-up, were included in this study. Patient with de novo stage IV disease, with bilateral breast cancer, receiving neoadjuvant therapy for breast cancer, or with previous malignancy history were excluded from this study (Supplementary Figure 1). Patient baseline clinical characteristics were extracted from Shanghai Jiaotong University Breast Cancer Database (SJTU-BCDB).

Pathological Assessment
Histopathological assessment and immunohistochemical (IHC) evaluation were conducted in the Department of Pathology, Ruijin Hospital, by at least two independent experienced pathologists.
Estrogen receptor (ER) positivity and progesterone receptor (PR) positivity were defined as 1% or more positive invasive tumor cells with nuclear staining (19). HER2 status was first determined by IHC staining and scored as 0, 1+, 2+, and 3+ according to the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guideline (20). Samples with HER2 IHC 2+ were further examined by fluorescence in situ hybridization (FISH). HER2 positivity was defined as HER2 IHC 3+ or FISH positive. Five breast cancer molecular subtypes were classified according to the 2013 St. Gallen breast cancer consensus (21)

Follow-Up and Disease Outcomes
Follow-up was accomplished annually by specialized breast cancer nurses in our center through outpatient medical history and/or phone calls. Recurrences in ipsilateral breast, chest wall, or regional LN (ipsilateral axillary, infra-and/or supraclavicular, or internal mammary LN) were considered LRR. DM included metastases to distant LN, bone, brain, liver, lung (including pleura and lymphangitic carcinomatosis), or others (including peritoneal, other organs not elsewhere classified, and skin not in the breast and chest wall). Patients with concurrent LRR and DM were categorized as DM as first recurrence event.
Recurrence-free interval (RFI) was defined as time from the date of breast cancer surgery to the date of first recurrence event. OS was defined as time from the date of breast cancer surgery to the date of death from any cause. Post-recurrence OS (PR-OS) was defined as the time from the date of first recurrence diagnosis to the time of death from any cause.

Statistical Analysis
Patients were categorized into three groups according to their recurrence status, i.e., recurrence-free, LRR, and DM groups. Descriptive characteristics of categorical variables were tested using chi-squared test or Fisher's exact test. Binary or multinomial logistic regression analysis was conducted to compare baseline clinicopathological features and adjuvant therapy among groups. Survival curves were plotted using the Kaplan-Meier method and compared between groups by logrank test. Multivariate Cox proportional-hazards regression analyses were performed to calculate hazard ratios (HRs) and 95% confidence interval (CI) for recurrence and survival. All analyses were performed using IBM SPSS 22.0 (IBM Inc., Armonk, USA). All reported p-values were two-sided, and p < 0.05 was considered statistically significant.

Factors Influencing Salvage Surgery for Locoregional Recurrence Patients
Forty out of 87 (46.0%) LRR patients received further salvage surgery. Table 3 summarizes the clinicopathological features associated with the reception of salvage surgery in LRR patients. Age at recurrence, primary tumor size, primary lymph node status, primary surgery of the breast and axilla, and LRR type significantly influenced the choice of surgery for LRR (p < 0.05; Table 3). Patients with IBTR received more salvage surgery as compared with LRR patients with chest wall recurrence or reginal LN recurrence (p < 0.001). Twenty-one out of 26 (80.8%) patients with IBTR received salvage surgery, all of whom received mastectomy with or without ALND. Only five patients with isolated IBTR did not receive surgery for LRR, including two patients refusing further treatment, two treated with endocrine therapy but not surgery due to advanced age, and one participating in a clinical trial of a new drug. Twelve out of 27 (44.4%) patients with chest wall recurrence received extended tumor excision, while seven out of 34 (20.6%) patients with regional LN recurrence received LN dissection surgery. Among 27 patients who did not receive surgery for regional LN recurrence, nine, 17, and one patients were with ALN recurrence, supraclavicular/infraclavicular LN recurrence, and internal mammary LN recurrence.
Multivariate analysis showed that primary tumor size (p = 0.039), primary surgery of the axilla (p = 0.006), and LRR type (p < 0.001) were factors that independently influenced the choice of surgery for LRR ( Table 4). Patients with smaller primary tumor size, primary SLNB, and IBTR had significantly higher probability to receive surgical treatment for LRR. Patients with regional LN recurrence were less likely to receive surgery for LRR than were patients with IBTR only (OR = 0.07, 95% CI 0.02-0.30, p < 0.001), while the probability of surgery for LRR was comparable between patients with chest wall recurrence and IBTR (OR = 0.36, 95% CI 0.09-1.47, p = 0.155).

Survival Outcome With Different Recurrence Events
The estimated 5-year OS was 80.7%, 50.3%, and 98.8% for patients with LRR, patients with DM, and recurrence-free patients, respectively (p < 0.001, Figure 1). Among the 87 patients with LRR, 26, 27, and 34 patients had IBTR, chest wall recurrence, and LN recurrence, respectively. During a median post-recurrence follow-up time of 21.3 (range: 1.0-77.5) months, 30 deaths were recorded. PR-OS curve is shown in Figure 2A.
Patients with LRR as first event had a significantly better PR-OS than those with DM (3-year PR-OS 75.0% vs. 37.1%; p < 0.001, Figure 2A).

DISCUSSION
In this cohort of 5,202 consecutive breast cancer patients, we showed that LRR after radical surgery in the modern era is relatively low. Clinicopathological factors, including large tumor size, positive lymph node status, and molecular subtype, were significantly associated with increased risk of LRR. Primary surgical treatment for breast or adjuvant chemotherapy or   radiotherapy also influenced the risk of LRR. Moreover, LRR patients had higher rates of receiving BCS or SLNB and not receiving adjuvant chemotherapy or radiotherapy compared with DM patients. Furthermore, we found that LRR types were related with salvage surgery choice after LRR. For patients receiving surgery after LRR, they could achieve an excellent outcome after recurrence. According to the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) overview, which included trials up to year 2000 evaluating the effects of radiotherapy, the 5-year LRR rate was 7% in patients after BCS and radiotherapy and 6% in patients after mastectomy (22). A reduction of LRR has been seen in the recent years with the improvement in imaging, earlier diagnosis, surgical planning, and adjuvant therapy for breast cancer patients (5). In our study, the 5-year LRR rate was 2.8% in the whole population: 3.8% in patients receiving BCS and 2.5% in patients receiving mastectomy, which were quite low compared with the established evidence. The low LRR rate highlights the effect of multiple changes in breast cancer management over the past two decades.
Several clinicopathological factors as well as treatment patterns were associated with LRR after surgery in early breast cancer patients. Not surprisingly, in our study, we found that large tumor size, positive LN status, and primary BCS were identified as independent risk factors for LRR, which was consistent with previous studies (23,24). Meanwhile, adjuvant chemotherapy and radiotherapy can effectively reduce the risk of LRR. Neoadjuvant chemotherapy was one of risk factors for local recurrence as reported by the EBCTCG meta-analysis (25), but neoadjuvant population was not included in our study. There was controversy in grouping patients when analyzing  the two populations together, since there is discordance of molecular biomarkers before and after neoadjuvant therapy, and the staging of patients will change after neoadjuvant therapy. Also, in neoadjuvant study, we usually use event-free survival to evaluate patients' outcome, which includes more information than recurrence-free interval that we evaluated in adjuvant studies. By reason of the foregoing, we excluded patients who received neoadjuvant therapy in this study, to make the evaluation standardized in the whole study population. We also found that LRR was a less common recurrence event, as either first recurrence event or subsequent recurrence event comparing with DM. Few studies directly compared the difference between patients with different first recurrence events. Our study demonstrated that LRR patients had higher rate of receiving primary BCS, primary SLNB, and lower rate of receiving adjuvant chemotherapy or radiotherapy, indicating that more effective systemic and local treatment should be evaluated to further reduce the rate of LRR.
In the modern era of breast cancer treatment, management of LRR breast cancer patients remained a big challenge due to lower LRR events, fewer high quality clinical evidence, and relatively hard to follow-up patients. For patients who developed IBTR after BCS, the current standard of care is further salvage surgery, including salvage mastectomy or repeat BCS (26), which can achieve 59% to 90.9% 5-year OS after salvage surgery (11,(27)(28)(29)(30). There is also another special consideration for patients with IBTR that whether it is "true recurrence" or "new primary," since new primaries should theoretically have a prognosis independent of the primary breast cancer. The rate of new primary breast cancer in patients with IBTR was 18%-58.9% in published studies (31)(32)(33)(34), also strengthening the reason for surgery of IBTR. For patients with isolated chest wall recurrence, full-thickness chest wall resection can be performed with excellent survival and low morbidity. In a recent systematic meta-analysis of 48 studies accounting for 1,305 patients who received full-thickness resection for chest wall recurrence, the mortality was consistently low (<1%), and 5-year OS was 40.8% (8). Axillary recurrence rates are rare, ranging of 1% to 3% after adequate management of primary disease (35,36). Salvage ALND was the first choice for selected patients and can be performed in 45.5% to 69.5% patients (37,38). Surgery of LRR might be encouraged in patients who can achieve R0 resection. In our study, salvage surgery was performed in 46.0% of LRR patients: 80.8% for IBTR, 44.4% for chest wall recurrence, and 20.6% for regional LN recurrence. Patients with smaller primary tumor, receiving primary SLNB, and LRR type were related with the choice of surgery after LRR. Although the post-LRR follow-up period is short, and there was selective bias in patients receiving salvage surgery, we do observe that patients receiving surgery for LRR achieved a better PR-OS, which emphasized the importance of surgery as part of multidisciplinary management of LRR patients.
Some limitations of this study exist. The data were collected retrospectively, which may have led to selection bias. The followup time is relatively short, and only a small number of LRR events were recorded, given that LRR was less common in clinical practice. The actual site of recurrence may influence the possibility of surgery for LRR lesions and were not analyzed in this study. Details of the recurrence including site and pathologic features of the recurrent lesion are not completely collected, and we cannot distinguish whether there is true recurrence or new primary breast cancer in patients with IBTR. Treatments of LRR out of surgery such as systemic therapy or radiotherapy and their impact on survival were not recorded or analyzed in this study. More comprehensive treatment data as well as longer follow-up are warranted to find the best management for LRR patients.

CONCLUSION
LRR rate was relatively low in the modern era of breast cancer treatment cohort. Large tumor size, positive lymph node status, and treatment strategies were associated with LRR. Moreover, LRR patients had a higher rate of receiving primary BCS or SLNB, and not receiving adjuvant chemotherapy or radiotherapy compared with DM patients. LRR patients treated with salvage surgery experienced excellent survival, indicating salvage surgery should play an important role in multidisciplinary treatment of LRR patients.

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 Ethical Committees of Ruijin Hospital, Shanghai Jiaotong University School of Medicine. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. HR, hazard ratio; CI, confidence interval; ER, estrogen receptor; ALND, axillary lymph node dissection; SLNB, sentinel lymph node biopsy; LRR, locoregional recurrence; IBTR, ipsilateral breast tumor recurrence; LNR, lymph node recurrence. *Tumor characteristics were from primary breast cancer. Patients with unknown lymph node status were not included in multivariate analysis.