Management of the axilla in postmenopausal patients with cN0 hormone receptor-positive/HER2-negative breast cancer treated with neoadjuvant endocrine therapy and its prognostic impact

To evaluate the differences in nodal positivity if the sentinel lymph node biopsy (SLNB) is performed before or after neoadjuvant endocrine therapy (NET) in breast cancer patients, and its impact on prognosis. A retrospective cohort study was performed in a single center including 91 postmenopausal cases with clinically node-negative and hormone receptor-positive/HER2-negative (HR + /HER2−) breast cancer, treated with NET and SLNB. SLNB was done pre-NET until 2014, and post-NET thereafter. Axillary lymph node dissection (ALND) was indicated only in SLNB macrometastasis, although in selected elderly patients, it was omitted. Kaplan–Meier survival curves were estimated in relation to the status of the axilla, and the differences assessed using the log-rank test. Between December 2006 and March 2022, SLNB was performed pre-NET in 14 cases and post-NET in 77. Both groups were similar in baseline tumor and patient characteristics. SLNB positivity was similar regardless of whether SLNB was performed before (5/14, 35.7%) or after NET (27/77, 37%), with 2/14 SLN macrometastases in the pre-NET cohort and 17/77 in the post-NET cohort. Only three patients (18.7%) with SLN macrometastasis had > 3 positive nodes following ALND. The 5-year overall survival and distant disease-free survival were 92.4% and 94.8%, respectively, with no significant differences according to SLNB status (p 0.5 and 0.8, respectively). SLN positivity did not differ according to its timing (before or after NET). Therefore, NET has no effect on lymph node clearance. Furthermore, the prognosis is good regardless of the axillary involvement. Therefore, factors other than axillary involvement may affect the prognosis in these patients.


Introduction
Hormone receptor-positive/human epidermal growth factor receptor 2-negative (HR + /HER2−) tumors are the most common type of breast cancer, accounting for approximately 75% of women over 50 years old [1]. In a recent review of Waks and Winer [2], the 5-year breast cancerspecific survival in stage I for HR + /HER2− tumors was > 99%, while for HER2 + was > 94% and for triple negatives was > 85%. Furthermore, Gianni et al [3] have found a significant correlation between HR + and low response to chemotherapy, in a large series of cases with breast cancer, where HR + patients achieve a complete pathological response between 5 and 6 times less than the rest. Therefore, this subgroup of tumors has a better prognosis than others [2], is less sensitive to chemotherapy [3,4], and requires different therapeutic approaches. In contrast, current axillary management guidelines for these patients does not differ from that applied in other more aggressive subtypes.
Several clinical trials have demonstrated that neoadjuvant endocrine therapy (NET) in postmenopausal patients can downsize HR + /HER2− tumors and facilitate breastconserving surgery [5]. Aromatase inhibitors are the treatment of choice [6]. However, axillary management strategies are underreported in these patients; indeed, they have been excluded not just from the large trials evaluating the feasibility of sentinel lymph node biopsy (SLNB) after preoperative therapy [7] but also from most of the ongoing clinical trials investigating regional nodal radiation after positive SLNB and neoadjuvant systemic therapy (NST). As a result, information on axillary management in the NET setting is lacking.
Neoadjuvant chemotherapy (NAC) has been shown to reduce the requirement for axillary lymph node dissection (ALND) in clinically node-positive (cN +) patients, since it achieves a complete pathological axillary response (pCR) in more than 60% of cases, especially for HER2positive and triple negative tumors [8,9]. By contrast, the incidence of axillary pCR after NET is much lower (< 10%) [10]. Moreover, in node-negative (cN0) disease, ALND rates also fell from 42% when performed before NAC to 12% following NAC [11], thanks to the downstaging of the undetectable lymph nodes involved. However, no data on SLNB before or after NET were reported.
The lack of significant differences in survival outcomes observed in patients aged ≥ 60 years with HR + / HER2− disease treated with or without ALND [12] suggests an opportunity to de-escalate treatment when patients have limited residual nodal disease. In fact, information from the National Cancer Database after examining patterns of axillary management in patients with cN0 HR + / HER2− disease revealed that those who had a positive SLNB after NET were less likely to complete ALND, thus favoring the adoption of less aggressive management strategies by this group [13]. Nevertheless, the long-term impact of omitting ALND in this setting remains unknown.
The efficacy and tolerability of NET have been widely reported, especially in elderly patients, and its superior cost-efficacy balance compared with NAC has been stressed [14]. However, it is not widely used in clinical practice [15]. Given the limited data available for axillary management after NET, the present study aimed to evaluate the differences in nodal positivity if the SLNB is performed before or after NET in patients with cN0 HR + /HER2− breast cancer, as well as its impact on adjuvant chemotherapy decisionmaking and prognosis.

Study design and patients
This retrospective cohort study of a prospectively maintained database included cN0 breast cancer patients treated with NET at Institut Català d'Oncologia-Hospital Universitari Bellvitge, between December 2006 and March 2022. In that period, 522 patients were identified as candidates for NST, and 91 were finally treated with NET (17.4%) (see Fig. 1). The Hospital Universitari Bellvitge Institutional Review Board approved the study (247/06), and all patients signed informed consent forms. The anonymization of all patient data was ensured, and their confidential information was protected in accordance with the national regulations.
Patients were required to meet all the following eligibility criteria: (1) postmenopausal women, or men; (2) tumor size > 2 cm or palpable breast tumor; (3) clinical and radiological cN0 at diagnosis (by axillary ultrasound); (4) invasive breast carcinoma in tumor core biopsy, estrogen receptor (ER) > 50% and HER2-negative tumors; (5) treated with NET; and (6) willingness to undergo surgery following NET. We excluded patients who had axillary lymph node (ALN) positive disease at diagnosis by fine-needle aspiration, treated with NAC, and HER2-positive tumors.

Treatments and cohorts
NET was indicated in patients with HR + /HER2− in which primary surgery could be less esthetic, trying to downstaging the tumor and thus improve surgery, knowing that this type of tumor does not respond well to neoadjuvant chemotherapy, or older patients in whom chemotherapy would ideally be avoided, to test endocrine sensitivity in vivo according to our multidisciplinary committee. Treatment consisted of 2.5 mg letrozole in women and 20 mg tamoxifen in men for 1 3 6-12 months, depending on response. Conservative breast surgery or mastectomy following NET was decided according to tumor response, tumor size, and predictable esthetic results.
Cases were distributed into two cohorts based on when they underwent SLNB before (pre-NET) and after NET (post-NET). SLNB was performed before NST from December 2006 to April 2014 and following NST thereafter according to our breast cancer guidelines. Peritumor injection of 3 mCi/mL 99 m-technetium radiocolloid (99 m-Tc) was administered for lymphatic mapping 24 h before the SLNB. When initial migration had not been observed on a previous gammagraphic map, we performed a second injection with 99 m-Tc the day before to improve migration. Radioactivity was detected with an intraoperative gamma probe (Europrobe; Britec, Sheffield, United Kingdom).
In the pathology department, the SLN and adipose tissue were fixed in 10% formaldehyde and completely embedded in paraffin, and processed in 3-mm sections, separated by 150 mm, three stained using hematoxylin and eosin and three stained using AE1/AE3 cytokeratin. Micrometastasis was defined as a small cluster of cells measuring 0.2-2 mm, and macrometastasis as > 2 mm. In both cohorts, deferred ALND was performed in patients with SLNB macrometastasis, except in selected elderly women in whom it was omitted in accordance with the Rudenstam et al. guidelines [12]. Patients with negative SLNs or with SLN micrometastasis did not undergo further treatment of the axilla.
Adjuvant chemotherapy was offered to patients with a postoperative endocrine prognostic index [16] (PEPI score) > 3 if they were aged < 80 years and had a good performance status (< 2). Most of the schemas included taxanes; however, therapy was adapted to each patient depending on their associated comorbidities prior to oncogeriatric screening. All patients received the same adjuvant endocrine therapy as in the neoadjuvant setting (2.5 mg letrozole, except in cases of progression during NET) to complete a minimum of five years. All patients had radiotherapy of the breast or chest wall according to our institutional guidelines after breast-conserving surgery and in cases with cT3-T4 tumors after mastectomy. Patients with SLNB-positive disease also received supraclavicular radiotherapy, including level III-ALN.

Variables
We recorded the following clinical and pathologic variables: age, tumor size before and after NET, histological type, Nottingham histological grade, lympho-vascular invasion, hormonal receptors status, proliferative index by Ki-67 staining, percentage of TILs (considered negative if < 5% of stromal TILs), treatment duration, type of breast and axillary surgery, SLN identification rate, SLN involvement, ALND rate, and any adjuvant treatments. Samples were divided into two surrogate subtypes according to PR and Ki-67 expression [17]: Luminal A-like, for progesterone receptor (PR) > 20%, and Ki-67 < 20%; and Luminal B-like (HER2−), for PR < 20% or Ki-67 > 20%. The Ki-67 proliferative index was grouped as > 30% or < 30% based on the results from a receiver operating characteristic curve (data not shown) at diagnosis and > 10% or < 10% after NET. Evaluation of clinical and radiological responses was grouped according to World Health Organization criteria, as complete response, partial response, no response, or progression [18]. Pathological complete response was considered when the pathologist observed 100% fibrosis replacing the invasive carcinoma and no axillary involvement, partial response if > 30% of fibrosis, and no response if < 30% fibrosis.

Follow-up
The patients were evaluated clinically and radiologically every three months before surgery. Radiological assessment included mammography and ultrasound of the breast and axilla in all cases, with magnetic resonance imaging reserved for selected cases. After surgery, patients were followed clinically every six months, with an annual mammography. Recurrences were defined as local when present in the ipsilateral breast or lymph nodes and systemic when they involved contralateral lymph nodes or other organs. Contralateral tumors were recorded separately.

Statistical analysis
Categorical variables are presented as the number of cases and proportions, whereas normally distributed continuous variables are presented as means and standard deviations (SD). Differences between the pre-NET and post-NET groups were analyzed by the chi-square test or Fisher's exact test for categorical variables and by the student t-test for continuous variables. The level of statistical significance was arbitrarily set at 5%. Statistical analyses were performed using IBM SPSS version 23.0 (IBM Corp., Armonk, NY, USA). Distant disease-free survival (DDFS) was defined as the time from diagnosis to distant recurrence, a second primary invasive cancer, or death from any cause. Breast cancer-specific survival (BCSS) was calculated from diagnosis until death from breast cancer, and overall survival (OS) was calculated from diagnosis until death from any cause. OS, BCSS, and DDFS were calculated at 5 years with their 95% confidence intervals (95%CIs). Kaplan-Meier survival curves of DDFS and OS were obtained in relation to the status of the axilla, which was grouped as ypN0 or ypN + , to assess the impact of affected lymph nodes in the context of HR + HER2− tumors, and the differences were computed by the log-rank test. Patients without events were censored at the time of the last follow-up.

Results
From December 2006 to March 2022, we included 89 patients who underwent NET and SLNB (88 women and 1 man); two of the patients had bilateral synchronous disease, so the final sample comprised 91 cases. The man included in the study was 93 years old and presented a cT4bN0 invasive ductal carcinoma at diagnosis with partial response after NET (ypN2ypN0). The PEPI score after NET was 1, so adjuvant CT could be safely avoided. After 13 months of follow-up, the patient is still alive and free of disease. SLNB was performed before NET in 14 cases (pre-NET cohort) and following NET in 77 (post-NET cohort). Table 1 shows the baseline characteristics of the patients and tumors, as well as the clinical-radiological response in the overall sample and comparing both cohorts pre-vs post-NET. There are no significant differences between both cohorts in the baseline description The seven cT1N0 cases included were due to bilateral tumors (n = 1), comorbidity associated with the diagnosis (n = 4), or COVID infection requiring surgical delay (n = 2). Mammographic evaluation of a breast tumor before and after NET, with partial response, is shown in Fig. 2.
Median months under NET were 7.6 (ranged 4.8-12.4) in the pre-NET group and 7.7 (ranged 4.1-13.6) in the post-NET group. SLNB identification rate was higher before NET although non-significant (p = 1): 100% versus 94.5% after NET (Table 2). In three of the four undetectable cases, ALND was omitted because the patients were elderly. SLNB positivity was similar whether performed before or after NET (35.7% versus 37%, respectively) and the rate of omission of ALND was similar in the two cohorts (85.7% pre-NET and 88.2% post-NET, p = 1). Note that only three of the 17 ALND performed (17.6%) had > 3 positive nodes in the final pathology report and that all these cases had more than one positive SLNB. Table 3 details the local and systemic adjuvant therapy and the results of pathological specimen evaluation. Interestingly, pathological complete response was observed in one patient with a cT2N0 luminal B-like tumor. Ki-67 staining after NET fell to < 10% in most patients (60.4%). The PEPI score was not evaluated until 2014, since then 20 out of 77 patients (26%) have had a PEPI score of ≥ 4; only four of these did not receive adjuvant chemotherapy due to low performance status or comorbidities.
After a mean follow-up of 50 months (standard deviation, 38; range 9-174 months), five recurrences and four deaths were reported (Table 3). A locoregional recurrence occurred in one patient with pN1mi disease (pre-NET cohort), contralateral breast cancer in one patient with ypN1a disease (post-NET cohort), and distant recurrences in three patients with pN0/ypN0 disease (one in the pre-NET cohort and two in the post-NET cohort). Of the four patients who died, only one was due to breast cancer. The 5-year OS was 92.4% (95%CI 84.5-100%), the BCSS was 98.3% (95%CI 94.9-100%), and the DDFS was 94.8% (95%CI 87.4-100%) in the overall sample (Fig. 3). When comparing DDFS and OS, no significant differences were found between positive and negative SLNB (Fig. 4). Long rank for DDFS was 0.07, p = 0.8, and for OS was 0.4, p = 0.5.

Discussion
In our sample of 91 cN0 breast tumors from 89 patients under neoadjuvant endocrine therapy, SLNB was positive in 35.7% of cases when performed pre-NET and in 37% of cases post-NET. The overlap of pre-and post-NET results indicates that endocrine therapy has no effect on subclinical SLN-positive disease. In contrast, previous studies by our team on SLNB before and after neoadjuvant chemotherapy observed that NAC reduces SLNB involvement threefold in cN0 patients (from 42% pre-NAC to 12% post-NAC), indicating that NAC provides a benefit in subclinical SLNpositive disease [11]. There are few studies evaluating SLNB in the setting of NET [19] but, based on the overlap of our results before and after NET, it appears that the effect on lymph nodes differs substantially between chemotherapy and endocrine therapy. Seeking to answer this question, Friedman et al [20] evaluated axillary downstaging in HR + / The proportion of micrometastases/macrometastases in SLNB and in ALND was higher before NET than after, Fig. 2 Mammographic evaluation of a breast tumor before and after neoadjuvant endocrine therapy (NET), with partial response. (a,b) Craniocaudal (CC) and oblique projection of breast cancer tumor at diagnosis with metallic marker inside (c,d) CC and oblique projection 6 months after NET although the sample is very small. These results are consistent with those previously published by our group [11]. NET decreases the proliferation of tumor cells as it is seen in the decrease in KI67, but rarely converts a positive axilla into a negative one after NET, neither after NAC in luminal cases as the ones reported in this series. Perhaps the effect of NST could be easier when the SLN has micrometastases at diagnosis, managing to clean the affected node, but it would be more difficult when there is a greater tumor burden, so the disease would persist in a higher proportion of cases with macrometastasis afterward of the NST, but this hypothesis cannot be proved.
Lymph node disease in HR + /Her2− tumors is not usually very extensive. In our study, only three patients with SLN macrometastases (17.6%) had > 3 positive nodes in the final pathology study. Similarly, Kantor et al. observed the presence of ≤ 3 positive nodes in surgery in 90% of patients with cN0 after NET [21]. These results suggest that most patients with cN0 disease treated with NET have a low nodal disease burden, and therefore, might be spared the morbidity associated with ALND. In the United States National Cancer Database, no significant differences in 5-year OS were observed when comparing patients with residual isolated tumor cells or micrometastasis with those who had no residual nodal disease after NET [22]. Our findings support these data, with DDFS and OS at 5 years unrelated to SLNB involvement and both exceeding 95%. Therefore, NET does not seem to carry the same prognostic significance as the residual nodal disease reported after NAC by Fisher et al [23].
Despite the behavioral differences between breast cancer surrogate molecular subtypes and the impact of different types of neoadjuvant systemic therapy (NAC or NET), axillary management after NET is currently extrapolated from NAC, meaning that patients who are cN0 at diagnosis and SLNB-positive after NET will undergo ALND [6]. Emerging data suggest that the residual nodal disease burden for cN0 HR + /HER2− after NET should be similar to that for cT1-T2N0 after primary surgery, and that this may be an opportunity to extrapolate management strategies used in the upfront surgery setting. Both ACOSOG Z0011 [24,25] and AMAROS studies [26,27] have shown that cN0 patients undergoing primary surgery with one or two positive SLNBs did not benefit from ALND in terms of DDFS and OS. Thus, it might be possible to consider de-escalating axillary management strategies after NET, subject to the results of further studies that are needed to standardize surgical management.
Regarding adjuvant systemic therapy in HR + / HER2− disease, one of the most important questions is who will benefit from chemotherapy and endocrine therapy. Nitz et al. observed that patients with ypN0-1 disease (≤ 3 lymph nodes) and low Ki-67 after a short NET can safely receive adjuvant ET alone [28]. Another tool for deciding the need for adjuvant chemotherapy is the PEPI score, based on residual tumor size and Ki-67 after NET, but the value for lymph node involvement is the same regardless of the number of positive nodes [16]. Interestingly, the POETIC trial demonstrated that Ki-67 measured two weeks after preoperative NET was able to predict survival outcomes, providing a basis for simple and inexpensive testing that could personalize adjuvant treatment in HR + /HER2− disease regardless of the axillary involvement [29]. Another useful test to predict the NET response and help us select the target patients is the OncotypeDX© Recurrence Score (RS) multigene panel [30]. In a recent review, they noted that breast cancers with low-intermediate RS on core biopsy are four times more likely to respond to NET than those with high-risk RS.
No randomized controlled trials have been reported on neoadjuvant endocrine treatment in man with breast cancer due to its low incidence. In clinical practice, the same management is offered to men than to women according to the TNM and molecular surrogate subtype [31]. In consequence, the male case reported in our series meet clinical criteria to receive NET and, in fact, his evolution has been like their counterpart female patients.
Limitations of the study include the retrospective design which has an inherent selection bias for those patients chosen for this approach versus initial surgery or neoadjuvant chemotherapy, so these results may be different prospectively or in a clinical trial. Furthermore, the small number of pre-NET SLNBs, however it provides an understanding of the low effect of NET on axillary tumor burden. Another limitation of our study is the relatively short follow-up (50 months) given that recurrences in HR + /HER2− disease may occur beyond 5 or 10 years from surgery. Despite these limitations, our study is one of the few to have exclusively analyzed the effect of NET on patients with cN0 disease at a single institution.
Another important consideration is that most patients undergoing NET were over the age of 70, and in these patients, survival depended more on other associated comorbidities than on breast cancer diagnosis (e.g., only  one out of four deaths was related to breast cancer). In fact, in 2006, a clinical trial showed that avoiding axillary clearance in clinically node-negative women aged ≥ 60 years receiving NET had the same prognosis with better early quality of life than ALND; therefore, even SLNB could have been avoided for this population with HR + / HER2− disease [12]. Nevertheless, since life expectancy has risen over the years, SLNB was performed in most of these patients, and if positive, ALND might be offered. According to all data collected related to NET, Kantor et al. proposed a new tailored approach to axillary management after NET [32] based on performing an axillary ultrasound evaluation at the time of diagnosis to enumerate the suspicious lymph nodes. Then, for candidates for NET who had cN0 disease at diagnosis and ≤ 3 SLNB-positive nodes following NET, those authors recommend routinely omitting ALND. For patients who have biopsy-proven cN1 disease with ≤ 2 suspicious nodes by axillary ultrasound following NET, they recommend performing SLNB in addition to the clipped biopsy-proven node, only omitting ALND when 1-2 positive nodes are found. This approach may reduce axillary morbidity in a significant proportion of patients treated with NET, but to date, there are no studies that have demonstrated its safety.
Further prospective studies on the axillary management approach after NET are needed to establish consensus management of the axilla in patients with HR + /HER2− disease. Our group has designed a phase 3 randomized multicenter clinical trial (Axillary Lymph Node Dissection Versus Axillary Radiotherapy in Patients with Positive Sentinel Node after Neoadjuvant Therapy: ADARNAT, NCT04889924), to demonstrate whether ALND can be safely omitted in patients with ≤ 3 SLN macrometastases following NAC or NET in order to reduce the morbidity associated with ALND (recruitment is ongoing).

Conclusion
More than 30% of cN0 HR + /HER2− breast cancer patients treated with NET have a positive SLNB, irrespective of whether assessment is performed before or after neoadjuvant therapy. This suggests that the therapy has little effect on the axilla. In addition, survival may not be related to axillary involvement in these patients, who tend to present with low residual axillary disease burden. Therefore, future efforts should focus on de-escalating ALND after NET.