Axillary lymph node dissection: Dead or still alive?

Abstract

Although sentinel lymph node biopsy is now the primary method of axillary staging and is therapeutic for patients with limited nodal disease, axillary lymph node dissection (ALND) is still necessary for staging in groups where sentinel lymph node biopsy has not been proven to be accurate and to maintain local control in those with a heavy axillary tumor burden. Additionally, newer approaches to systemic therapy tailored to risk level sometimes necessitate knowledge of the number of involved axillary nodes which can only be obtained with ALND. Ongoing trials will address whether there are additional circumstances where radiotherapy can replace ALND.

Highlights

• •Axillary surgery in BC has been de-escalated for a significant group of patients, but an ALND is still necessary for some
• •ALND remains necessary for staging in patients where SLNB has not been demonstrated to be accurate
• •This includes patients with clinically palpable lymph nodes and patients with cT4 or cN2-3 disease
• •ALND is required for local control in patients with a heavy axillary tumor burden and recurrent axillary disease
• •Use of newer systemic therapy may require complete axillary staging with ALND to determine eligibility for use.

1. Introduction

Management of the axillary nodes in breast cancer has evolved rapidly in the modern era. Axillary lymph node dissection (ALND), first described by Dr. William S. Halstead in 1894 [1] was historically the standard of care for all breast cancer patients with the aim of providing both local control and improving overall survival (OS). However, the recognition that ALND is associated with significant morbidity, including high rates of lymphedema and worse quality of life [[2], [3], [4], [5]], coupled with the increased detection of small cancers with a lower likelihood of nodal metastases in the screening mammography era, led to interest in alternative approaches to axillary staging and management.

The idea that axillary surgery could be eliminated in patients with breast cancer was first formally demonstrated in NSABP B-04, a trial initiated in 1971 which randomized clinically node-negative patients to radical mastectomy, simple mastectomy with nodal radiation therapy (RT), or simple mastectomy alone. Clinically node-positive patients were randomized to radical mastectomy or simple mastectomy with nodal RT [6]. In all patients, including those with clinically positive nodes, there was no difference in OS for patients who had no axillary surgery. In clinically node-negative patients, although 44.6% in the radical mastectomy arm had nodal metastases, only 17.8% in the simple mastectomy arm developed isolated nodal recurrence [7]. Although this trial did not result in the abandonment of ALND due to the importance of nodal status in selecting patients for adjuvant chemotherapy during that era, it suggested that ALND was a staging, rather than a therapeutic, procedure, and that radiotherapy was an alternative method of obtaining local control.

The idea of de-escalating axillary surgery was advanced with the application of sentinel lymph node biopsy (SLNB) to breast cancer staging by Giuliano and colleagues [8], and SLNB is now the standard of care for staging the clinically node-negative axilla. Subsequently, randomized trials demonstrated that SLNB with or without radiotherapy [2,[9], [10], [11]] provided equivalent local control and survival to ALND in patients with micrometastases or macrometastases in 1–2 sentinel lymph nodes (SLNs). More recently, SLNB after neoadjuvant chemotherapy (NAC) has been shown to accurately stage the axilla in patients who are clinically node negative prior to NAC as well as in those with nodal metastases who downstage in response to NAC, with local control and survival outcomes comparable to those seen in patients undergoing ALND [[12], [13], [14], [15], [16]]. This body of evidence, coupled with the fact that decisions regarding adjuvant systemic therapy are increasingly made on the basis of receptor status and genomic profiling rather than the presence or absence of nodal metastases or the number of involved nodes, has led some to question whether ALND is ever indicated in the management of primary breast cancer.

Current indications for ALND can be grouped into 3 categories; patients in whom evidence that SLNB accurately stages the axilla is lacking, patient subsets in which SLNB alone or with RT is not proven to result in local control or survival equivalent to that seen with ALND, and an emerging group of patients for whom knowledge of the extent of nodal involvement alters the choice of systemic therapy. This article will review these populations and describe ongoing clinical trials designed to address these issues.

2. Indications for ALND for accurate axillary staging

SLNB is widely accepted as an accurate technique to stage the axilla; however, its accuracy has not been demonstrated in all patients with breast cancer. Current National Comprehensive Cancer Network guidelines support the use of SLNB for axillary staging in patients who are clinically node negative by physical exam, have ≤2 suspicious nodes on imaging, and are undergoing upfront surgery [17]. These recommendations are supported by a high level of evidence gained from randomized controlled trials which enrolled clinically node-negative patients with T1-T2 tumors and demonstrated false-negative rates (FNRs) of <10% when SLNB was compared to ALND as a staging procedure [10,18,19].

ALND remains indicated in those patients undergoing upfront surgery who do not meet the inclusion criteria for trials validating the accuracy of SLNB. Although a small number of T3 patients were included in NSABP B-32 and the Milan sentinel node trial, the exact number of these patients is unclear, and they were not analyzed separately [18,20]. Patients undergoing upfront surgery for T4 tumors were not eligible for the multi-center randomized controlled trials of SLNB. In spite of this, SLNB appears to be performed frequently in patients with cT3 and cT4 tumors. A retrospective study utilizing the National Cancer Database (NCDB) examined patients with cT3 and non-inflammatory cT4 disease who had 1-2 positive SLNs following mastectomy. In this cohort of 1917 patients, 60% underwent ALND and 40% SLNB alone, with 58% of ALND patients and 56% of SLNB-alone patients also receiving RT [21]. The 10-year estimated OS was significantly better in those who underwent ALND compared to SLNB (59% versus 45%, p < 0.01), and this benefit persisted in those who received RT (63.5% ALND plus RT versus 53% ALND alone, versus 49% SLNB plus RT versus 39.5% SLNB alone). This difference in survival may be explained by the higher nodal disease burden observed in cT3 and non-inflammatory cT4 patients, with 47–59% of cT3 and non-inflammatory cT4 patients having positive axillary lymph nodes versus 33–42% of cT1-2 patients [22], and illustrates the fact that in patients with a heavy axillary tumor burden, both ALND and radiotherapy may be necessary for optimal disease control. Additionally, NAC is commonly indicated for patients with high T stage, and in particular, those with T4 disease, so meaningful data on the accuracy of SLNB in the upfront surgical setting are unlikely to be generated for this population. Based on the absence of data indicating the accuracy of SLNB in patients with T4 tumors undergoing upfront surgery, this remains an indication for ALND. While it is likely that SLNB is accurate for smaller T3 tumors, caution should be exercised in the use of SLNB for staging the axilla in very large T3 tumors.

Patients with clinically positive (palpable) nodes undergoing upfront surgery were also excluded from the multi-centered randomized control trials examining the accuracy of SLNB due to the concern that palpable adenopathy is associated with a heavy axillary tumor burden. An analysis of the 189 patients with clinically positive axillary nodes in NSABP B-18 demonstrated that 28% of patients had 4-9 positive nodes and 23% had ≥10 positive nodes [23]. This tumor burden is substantially different than what was present in clinically node-negative patients in the ACOSOG Z0011 and AMAROS trials who were randomized to ALND, where only 21% and 8%, respectively, were found to have involvement of >3 additional nodes at ALND [2,10]. At this time, palpable adenopathy remains an indication for ALND in patients undergoing upfront surgery. Unnecessary ALND can be minimized with the use of NAC in human epidermal growth factor receptor 2 (HER2) positive and triple negative cancers, while the decision to undertake surgery first versus NAC in hormone receptor positive, HER2 negative patients requires consideration of multiple factors.

ALND also remains necessary for accurate staging in the post-neoadjuvant setting in a subset of patients. In patients who have received NAC, the performance characteristics of SLNB vary with the presenting tumor stage [14,16,[24], [25], [26], [27]]. In patients with clinical T1-3, N0 disease pre-treatment, SLN identification rates, FNRs, and rates of local recurrence mirror those seen in the upfront surgical setting, and SLNB without ALND as the initial procedure for staging the axilla is standard practice. However, ALND remains necessary to stage the axilla in patients with pre-treatment T4 disease, even when an excellent response to NAC is seen, as the accuracy of SLNB in these patients has not been demonstrated, as discussed previously. The ACOSOG Z1071 trial of 701 patients undergoing SLNB after NAC included 33 patients (4.7%) with non-inflammatory T4 tumors; however, a separate analysis of this group was not performed [25]. Other multi-institutional prospective trials excluded all cT4 patients [14,16,24] or did not specify their inclusion [16]. Retrospective data are also limited by very small numbers. The largest cohort, described by Shen et al. included 12 cT4 patients out of 69 total cT1-4N0-3 patients who received NAC, and while the FNR in patients with T4 disease was not separately reported, it was noted that 33.3% of patients with T4 tumors did not localize with lymphatic mapping, a rate significantly higher than the 1.7% in the remainder of their cohort (p = 0.002) [28]. Additionally, the overall FNR in this study was an unacceptable 25%

Determining which patients who downstage from cN1 to cN0 after NAC require ALND to accurately stage the axilla is dependent upon the ability to perform SLNB with a low FNR. The accuracy of SLNB varies based upon the mapping technique used and number of SLNs retrieved. Four multi-institutional randomized controlled trials have demonstrated that in patients presenting with nodal metastases who downstage to cN0 after NAC, the overall FNR of SLNB ranged from 11.9 to 14.2%, higher than the 10% rate accepted for upfront surgery. When SLNB was performed with dual tracer localization, the FNR was decreased to 5.2–10.8%, and to 4.9–9.1% with the retrieval of ≥3 SLNs [14,16,24,25]. At institutions where clips are routinely placed after axillary lymph node biopsy, targeted removal of the clipped node in addition to the SLNs is a strategy that improves the FNR of SLNB to 1.4–6.8% [29,30]. When 3 or more SLNs or a clipped node and at least 1 additional SLN are retrieved, omitting routine ALND for staging purposes is appropriate. However, when only 1–2 SLNs are identified without obtaining a previously positive and clipped node, FNRs of SLNB for staging are unacceptably high at 18–24% [16,24,25], and ALND is required.

Controversy exists over the necessity of ALND in patients who continue to have palpable nodes at the completion of NAC. Of the 4 randomized controlled trials validating SLNB after NAC, 2 did not specify post-therapy clinical node status [14,24] and 1 mandated a clinically normal axilla to be eligible for SLNB [16]. The ACOSOG Z1071 trial of 701 patients undergoing SLNB after NAC included 88 patients (12.6%) with palpably abnormal lymph nodes post-treatment and demonstrated a 19.2% FNR in these patients [25]. In a retrospective study, Alvarado et al. reported an FNR of 33.3% in patients with persistent clinical nodal disease post-NAC [31]. In general, there is no role for SLNB after NAC in patients with persistent adenopathy that is clearly malignant. In patients with small, mobile nodes which do not clearly contain cancer, SLNB is a reasonable approach; however, the standard of care for patients with persistent lymphadenopathy suspicious for metastatic disease after NAC remains ALND.

ALND is indicated for staging as well as local control in almost all circumstances for patients with cN2-3 disease. While the vast majority of patients presenting today with cN2-3 disease, will be recommended for NAC, if cN2 or cN3 patients are undergoing upfront surgery, ALND is required for reasons discussed earlier regarding lack of proven accuracy of SLNB in patients with clinically node-positive disease who undergo upfront surgery. However, unlike cN1 patients who downstage to cN0 with NAC where SLNB is proven to be accurate with the caveats discussed above, patients with cN2 or cN3 disease at presentation, regardless of clinical response to treatment, require ALND to accurately stage the axilla. This recommendation is supported by the paucity of data examining the accuracy of SLNB with cN2-3 disease. Although the multicenter randomized control trials discussed above all included patients with cN2 disease, these patients made up only 2.4–6.0% of the randomized patients [14,16,24,25]. The FNR of SLNB for cN2 disease was reported by ACOSOG Z1071 as 0% based on 26 patients, and 0% in 4 patients in the SN FNAC trial. Data on this population were not reported in either the SENTINA or GANEA 2 trials. A single-institution prospective observational study of 51 patients who received NAC reported an FNR for SLNB of 38.9% in 23 cN2-3 patients, significantly higher than the 5.9% FNR observed in cN1patients [32]. Additionally, a large retrospective study utilizing the NCDB analyzed 1827 patients with cN3 disease and demonstrated that omission of ALND from multimodal therapy was associated with a worse OS, with a hazard ratio of 2.20 [33]. Given the absence of prospective data demonstrating accuracy of SLNB in cN2-3 patients and concerns of oncologic safety of omission of surgical axillary clearance, patients who present with cN2-3 disease, regardless of response to chemotherapy, require ALND.

ALND also remains standard of care for all patients with inflammatory breast cancer, regardless of the response to NAC. Patients with inflammatory breast cancer have been consistently excluded in studies examining the accuracy and oncologic safety of SLNB. Although there has been a national trend toward non-guideline concordant utilization of SLNB in patients with inflammatory breast cancer [34], eliminating ALND for staging is not supported by any evidence. Single-institution prospective feasibility studies that examined SLNB in patients with inflammatory breast cancer, which included between 8 and 20 patients per study, reported failure to detect an SLN in 20–75% of patients, with FNRs ranging from 0 to 25% [[35], [36], [37]]. The 2 studies reporting identification of a SLN in 75–80% of patients were those with the highest FNRs, ranging from 18 to 25% [36,37]. At this time, SLNB is contraindicated outside of the clinical trial setting given the absence of evidence demonstrating accuracy of staging and oncologic safety of eliminating complete axillary clearance in inflammatory breast cancer. Indications for ALND for staging purposes are summarized in Table 1.

Table 1.

Indications for axillary lymph node dissection.

Current Indications for Axillary Lymph Node Dissection
Accurate axillary staging
Upfront surgery or post-NAC with palpable lymphadenopathy
Clinically evident N2–N3 disease
cT3-T4 tumor and upfront surgery
Inflammatory breast cancer
Local control of axillary disease
≥3 positive SLNs and upfront surgery (BCT or mastectomy)
Upfront mastectomy with metastases in 1–2 SLNs and no indication for PMRT
Any residual axillary disease on SLNB after NAC
Ipsilateral axillary recurrence
Lymph node status to direct systemic therapy
Subset of patients whose benefit from CDK4/6 inhibitor therapy is dependent upon number of positive axillary lymph nodes and treatment may be altered by complete axillary staging

Abbreviations: NAC, neoadjuvant chemotherapy; SLN, sentinel lymph node; BCT, breast-conserving therapy; PMRT, postmastectomy radiation therapy; SLNB, sentinel lymph node biopsy.

3. ALND for local control

Advances in breast cancer care including improved RT, and systemic and hormonal therapy options have played a significant role in the ability to de-escalate axillary surgery while maintaining local control in the upfront surgery setting. Multiple randomized controlled trials have demonstrated that in patients with 1–2 positive SLNs undergoing breast-conserving surgery, the residual axillary tumor burden is low, with only 13–33% of patients having positive non-SLNs and 2–8% of patients having ≥3 positive non-SLNs [2,9,10,18]. In these studies where whole-breast RT, with or without nodal RT, and systemic therapy were given, ALND did not improve locoregional recurrence or OS at 10 years compared to patients randomized to SLNB alone [2,9,10,18].

In contrast, the presence of axillary metastases to ≥3 SLNs in patients undergoing upfront surgery is an indication for ALND. The tumor burden in these patients is higher, reported at 72% in a retrospective review by Dengel et al. who analyzed the extent of axillary disease burden in consecutive patients who fell outside of Z0011 criteria by having either ≥3 positive SLNs, or who had extracapsular extension and underwent ALND [38]. The precise tumor burden which cannot be controlled by RT is uncertain, but NSABP B-04 demonstrated that in patients with clinically positive nodes after 10 years of follow-up, regional nodal recurrence was seen in 1.0% of those undergoing ALND and 11.9% of those treated with RT [7].

The indications for ALND in patients undergoing mastectomy are less clear. It has been thought that whole-breast RT, which is a standard part of breast-conserving therapy (BCT), is important for local control when ALND is not performed and SLN metastases are present because the breast tangents cover some of the low axilla, even when node field RT is not administered. Mastectomy patients were not eligible for ACOSOG Z0011 and made up only 17% and 16% of patients in AMAROS and OTOASOAR, respectively, so in total, only 151 patients were treated with mastectomy and nodal RT in these studies, and they were not analyzed separately [2,10,11]. However, there is no reason to believe the efficacy of nodal RT varies based upon the operation performed on the breast, so in patients undergoing mastectomy who have an indication for postmastectomy RT (PMRT) based on the presence of metastases in 1 or 2 SLNs and primary tumor characteristics, ALND can be avoided. For those with metastases in 3 or more SLNs, ALND should be performed as it is in patients undergoing BCT. ALND is also indicated when the need for PMRT is not clear based on the presence of 1 or 2 nodal metastases. The frequency with which ALND will be required for this indication will vary with the criteria used for administering PMRT. If the presence of even a single nodal metastases is considered an indication for PMRT, only patients with metastases in 3 or more SLNs will require ALND. In a retrospective study from Memorial Sloan Kettering Cancer Center (New York, NY, USA) of 2207 consecutive T1-3cN0 patients undergoing upfront mastectomy, only 49 (2%) were found to have metastases in 3 or more SLNs, necessitating ALND [39]. The goal should be to minimize the use of both ALND and RT in patients with 1 or 2 nodal metastases, and this may require deferring the decision for ALND until complete pathology is available, rather than basing it upon the finding of nodal metastases on frozen section. Further insight into the importance of RT in maintaining local control after SLNB alone for positive SLNs will come from the POSNOC and SENOMAC trials, discussed later in this article.

Following NAC, ALND remains standard of care to provide local control in patients with positive SLNs. While the oncologic safety of omitting ALND in patients with metastases in 1–2 SLNs has been demonstrated in the upfront surgery setting [2,9,40], these findings have not been confirmed in patients receiving NAC. Findings from studies completed in the upfront surgery setting cannot be applied to patients who have undergone NAC, as the burden of additional axillary disease differs (Table 2). In the upfront surgical setting, micrometastases and isolated tumor cells (ITCs) in the SLN are associated with additional positive lymph nodes in 8–18% of cases, and Galimberti et al. demonstrated no advantage for patients randomized to ALND compared to SLNB alone in a trial limited to those with micrometastatic disease. In contrast, patients with micrometastatic disease as the largest tumor deposit on SLNB after NAC have additional positive non-SLNs in 38–64% of cases, a rate which does not differ significantly from the 59–62% rate of additional disease identified when macrometastases were present in the SLN [41,42]. We hypothesize that micrometastases after NAC may not be true micrometastases, but rather are macrometastases that had an incomplete response to therapy. This hypothesis is supported by the fact that the presence of both ITCs and micrometastases in the axilla after NAC is associated with worse disease-free survival (DFS) and OS than is seen in patients with no tumor in the nodes. Wong et al. demonstrated in both an institutional cohort and patients from the NCDB that 5-year DFS decreased from 88% in ypN0 patients to 74% in ypN0 [i+] patients and 75% in ypN1mi patients [43]. The impact of leaving behind the higher burden of axillary disease post-NAC on both local control and survival remains unclear, and ALND remains standard of care for patients with residual axillary disease following NAC.

Table 2.

Burden of non-SLN disease in studies with patients who underwent SLNB followed by completion ALND.

Study	Year	Largest tumor deposit on SLNB	Number of patients	Percent of patients with additional positive non-SLNs
Upfront Surgery
Giuliano et al.	2010	Any positive SLNa	365	27%
		Micrometastasis	137	10%
Galimberti et al.	2013	Micrometastasis	447	13%
Cserni et al. [44]	2008	Micrometastasis	435	18.2%
		Isolated tumor cells	82	8.5%
Donker et al.	2014	Any positive SLNa	744	33%
Savolt et al.	2017	Any positive SLNa	244	38.5%
NAC
Classe et al.	2019	Any positive SLNa	141	41.8%
Moo et al.	2018	Macrometastasis	121	62%
		Micrometastasis	44	64%
		Isolated tumor cells	6	17%
Sanders et al.	2022	Macrometastasis	201	58.6%
		Micrometastasis	26	38.5%
Boileau et al.	2015	Macrometastasis	61	56%
		Micrometastasis	8	38%
		Isolated tumor cells	7	57%

Abbreviations: ALND, axillary lymph node dissection; SLN, sentinel lymph node; SLNB, sentinel lymph node biopsy; NAC, neoadjuvant chemotherapy.

^aAny positive SLN defined as any SLN with macrometastases, micrometastases, or isolated tumor cells.

Axillary recurrence remains an indication for ALND to provide local control. Multimodal therapy including complete axillary clearance with ALND in addition to systemic and RT has been shown in retrospective, single-institution studies to improve distant DFS and OS as compared to patients in whom ALND was not performed. In a retrospective study of 220 patients with an isolated axillary recurrence, Konkin et al. found that patients undergoing treatment with ALND plus RT had a 63% 5-year OS, as compared to 35% and 38% in those treated with ALND alone or RT alone [45]. In a study of 44 patients with an axillary recurrence, Newman et al. found that achieving complete surgical control of the axillary disease reduced the rate of distant recurrence to 41% compared to 65% in those patients in whom axillary surgery was not performed [46]. These data support the continued need for multimodal therapy including completion ALND in patients with an axillary recurrence. Indications for ALND to maintain local control are summarized in Table 1.

4. ALND to guide systemic therapy

Historically, recommendations for systemic therapy were based on the presence or absence of axillary nodal metastases and, in some cases, the number of nodal metastases, making ALND an essential part of the management of patients with carcinoma. As previously discussed, SLNB reliably identifies nodal metastases, and with advances in our understanding of tumor biology, decisions regarding the need for systemic therapy and the type of systemic therapy are now often based on hormone receptor status, HER2 expression, and genomic signatures such as the 21-gene recurrence score [17]. For patients with HER2 overexpressing breast cancer and triple negative breast cancer, chemotherapy is indicated for all but the smallest cancers, regardless of nodal status. While nodal status may influence the chemotherapy backbone, use of dual HER2 blockade, or administration of pembrolizimab, the number of nodal metastases is not a determinant of choice of therapy, particularly since a majority of these patients receive NAC. In patients with hormone receptor positive disease, 2 recent clinical trials, RxPONDER and monarchE, have called into question the use of SLNB alone for axillary staging, as the extent of axillary disease was used to determine patient eligibility for omission of chemotherapy in the case of RxPONDER, or the addition of abemaciclib to endocrine therapy in the case of monarchE [47,48].

The ability of a genomic assay to predict the benefit of chemotherapy in hormone receptor positive breast cancer was first described in node-negative patients [49,50], and the presence of nodal macrometastases remained an indication for systemic chemotherapy until recently. The RxPONDER trial was designed to address the ability of the 21-gene recurrence score to predict survival benefit from chemotherapy in patients with 1–3 positive lymph nodes (N1) by randomizing 5083 women with T1-T2 tumors and a Recurrence Score ≤25 to endocrine therapy alone or chemotherapy in addition to endocrine therapy [48]. In premenopausal women, a significant 5-year DFS improvement was noted in the chemotherapy group, while no survival benefit for the addition of chemotherapy to endocrine therapy was seen in postmenopausal women. The finding of lack of chemotherapy benefit in postmenopausal women raised the question of whether ALND should be performed in those with 1–2 SLN metastases to determine the number of involved nodes since study eligibility was limited to those with 1–3 nodal metastases. While this is a reasonable question, several lines of evidence suggest that ALND is not needed for this purpose. Patients were randomized to RxPONDER after surgery, and 37.4% of the cohort was staged with SLNB alone. No trend toward a benefit for chemotherapy was seen in the SLN group compared to the ALND group and, similarly, no trend toward chemotherapy benefit was seen based on the number of involved nodes [48]. In addition, a substantial body of evidence indicates that in patients without palpable adenopathy, with or without a negative axillary ultrasound, the likelihood of involvement of 4 or more axillary nodes is low. McCartan et al. retrospectively studied 5142 patients with T1-3 tumors and a normal axillary clinical exam, and found that only 6% had involvement of more than 2 axillary nodes [51]. Lee et al. reporting a cohort of postmenopausal women with T1-2 disease who were clinically node negative, found that of 2639 with hormone receptor positive, HER2 negative disease, 3.5% had N2-3 disease [52], and Pilewskie et al. observed a 1% incidence of N2-3 disease in 1786 postmenopausal hormone receptor positive/HER2 negative patients with cT1-2N0 cancers [53]. Concerns about missing a heavy nodal tumor burden can be further alleviated with the use of axillary ultrasound as illustrated in the initial reports of the SOUND and INSEMA trials. In the SOUND trial, patients with T1 tumors and a normal axillary ultrasound were randomized to axillary surgery versus observation, and in the surgery arm, only 0.5% of patients had 4 or more involved nodes (Virgilio Sacchini, personal communication). The INSEMA trial included T1 and T2 patients with the same randomization, and 1.3% of those in the surgery arm were found to have metastases in 3 or more nodes [54]. Based on these studies, concerns about the need for ALND to determine the extent of nodal involvement in this patient group do not seem to be warranted.

A potentially more compelling argument for ALND can be made based on the results of the monarchE trial [47], a study which randomized 5637 “high-risk” hormone receptor positive/HER2 negative patients to receive adjuvant endocrine therapy alone versus the CDK4/6 inhibitor abemaciclib for 2 years in addition to endocrine therapy. Patients were defined as high risk based on the presence of 4 or more positive axillary lymph nodes or 1–3 positive nodes with additional high-risk features (tumor ≥5 cm, histologic grade 3, or Ki67 ≥ 20%). In these high-risk patients, the addition of abemaciclib to adjuvant endocrine therapy conferred a significant DFS benefit, with those receiving abemaciclib having a 2-year invasive DFS of 92.2% versus 88.7% in those who received endocrine therapy alone (p = 0.01).

Unlike the question posed by the RxPONDER results, which is one of identifying patients meeting the study eligibility criteria, since it is unknown if those with involvement of 4 or more nodes and RS ≤ 25 benefit from chemotherapy, the monarchE trial demonstrates a DFS benefit in the study population, making their identification somewhat more critical. How frequently patients with 1 or 2 involved SLNs with tumors less than 5 cm in size which are grade 1–2 and have a Ki67 < 20 have involvement of 4 or more lymph nodes has not been studied to the knowledge of the authors, but the data on extent of nodal involvement in patients with a clinically negative axilla discussed above would suggest that it is uncommon. Lee et al. [52] performed a multivariable analysis of their 2639-patient group and found that younger postmenopausal age (<65 years), lymphovascular invasion, multifocal or multicentric tumor, and T size greater than 2 cm were significantly associated with pN2-3 disease, suggesting that patients with 1–2 SLN metastases and these features might benefit from ALND to determine their eligibility for treatment with abemaciclib, but this approach would be justified only in this subset of patients.

5. Future considerations

Efforts to minimize the number of patients who continue to require ALND as part of their breast cancer care are ongoing. In the upfront surgery setting, 2 clinical trials are examining the axillary management of patients who have ≤2 positive SLNs, and who have cT3 disease or are undergoing mastectomy. The SENOMAC trial is a multicenter randomized controlled trial examining the oncologic safety of omitting ALND in patients with 1–2 positive SLNs using a broader cohort of patients than was included in ACOSOG Z0011 or AMAROS. The SENOMAC trial is enrolling cT1-3 cN0 patients undergoing either BCT or mastectomy and SLNB [55]. Those patients with ≤2 positive SLNs are randomized to either ALND or no further axillary surgery; adjuvant chemotherapy and RT are given per guidelines at that institution, with the mandate that RT not be altered based on randomization results. If significant numbers of patients undergoing mastectomy without PMRT are included in this study, it has the potential to address the question of whether RT is essential for local control in patients with positive SLNs not undergoing ALND.

The POSNOC trial is a multicenter, randomized, non-inferiority trial examining the oncologic safety of omitting axillary RT in patients similar to those who met criteria for ACOSOG Z0011. The POSNOC trial enrolled patients with cT1-2N0 breast cancer undergoing up-front lumpectomy or mastectomy and were found to have 1–2 positive SLNs [56]. The control arm of the study was either axillary RT or ALND, and the experimental arm had no additional axillary treatment. Adjuvant systemic therapy was administered in both arms. The primary study end point is 5-year axillary recurrence, and recruitment to this study is complete.

Additionally, there are 2 randomized controlled trials assessing the oncologic safety of omitting ALND in patients with positive SLNs following NAC. The Alliance A0011202 trial enrolled cT1-3N1 patients who completed NAC and had at least 1 SLN with macrometastatic disease, randomizing these patients to ALND and nodal RT versus no further axillary surgery and nodal RT [57]. The study's primary endpoint is recurrence-free survival, and enrollment has been completed. The TAXIS European multicenter randomized controlled trial addresses a similar question, and is currently enrolling T1-3N1 patients who, following NAC, have any level of residual cancer, including ITCs and micrometastases in the SLNs. The randomization is to ALND or what is termed “Targeted Axillary Surgery”, which includes removal of SLNs, the clipped node, and any palpably abnormal nodes so that the axilla is rendered grossly disease free [58]. The study's primary endpoint is DFS. The results of both trials have the potential to have a significant impact upon the current guidelines recommending ALND for local control in the post-neoadjuvant setting.

6. Conclusions

Significant advancements in the de-escalation of axillary surgery, supported by a high level of evidence, have occurred in recent decades. These studies have broadened the use of SLNB for axillary staging and for therapy in patients with a limited axillary tumor burden, resulting in fewer indications for ALND and minimizing the morbidity of treatment for patients. However, there are subsets of patients in whom ALND remains necessary as discussed in this article and summarized in Table 1. ALND to stage the axilla remains standard of care in patients with cT3-4 tumors undergoing upfront surgery, inflammatory breast cancer, clinically positive nodes (both in the upfront surgery setting or following NAC), and cN2-3 disease, regardless of response to NAC. Local control must still be obtained surgically with ALND in those with a high burden of axillary disease not yet proven to be adequately controlled by either systemic therapy or RT, including patients with ≥3 positive SLNs undergoing upfront surgery, any residual nodal tumor following NAC, inflammatory breast cancer, and axillary recurrence. Additionally, a select group of patients will require ALND to assess the precise extent of nodal involvement to determine eligibility for adding the CDK4/6 inhibitor abemaciclib to their endocrine therapy. While continued efforts are ongoing to determine which patients may be spared the morbidity of an ALND, there remains a significant portion of patients in whom ALND cannot yet be eliminated.

Funding

The preparation of this study was funded in part by NIH/NCI Cancer Center Support Grant No. P30 CA007848 to Memorial Sloan Kettering Cancer Center.

Ethical approval

Not required.

Declaration of competing interest

All authors have no conflict of interest disclosures to report.