Introduction
Lymph node micrometastases (LNMM) was defined as the presence of metastases no larger than 2 mm in the lymph nodes, which was firstly proposed by Huvos et al. in 1971 [
1]. In 2002, the 6th edition of the American Joint Committee on Cancer (AJCC) Manual for Staging of Cancer based on lymph node involvement status, classified N staging into macrometastases (metastases lager than 2 mm), micrometastases (N1mi, metastases 0.2–2 mm in size), and isolated tumor cells (ITC, single tumor cells or small clusters of cells not larger than 0.2 mm, pN0(i +)) [
2,
3]. After that, Patani et al. [
4,
5] analyzed relevant literature on breast cancer LNMM, found that 12 of them (2000–2006) supported LNMM with independent prognostic significance and the prognosis is worse than that of lymph nodes without metastasis. While a recent multi-center cohort study from Sweden [
6] showed that lymph node micrometastases were associated with significantly lower 10-year breast cancer-specific survival (BCSS) and overall survival (OS) rates compared with lymph node-negative cases, while outcomes were similar to those of lymph node macrometastases, which may be associated with inadequate systemic treatment. Although the conclusions of different studies are inconsistent, LNMM is still considered to be an important prognostic factor of breast cancer.
With the progress of comprehensive treatments of breast cancer, surgical treatment has gradually become more precise and less invasive. Hence, the local treatments of patients with LNMM have attracted more attention to further improve. Previous large clinical studies [
7‐
11] have attempted to address the question of whether axillary lymph node dissection (ALND) can be safely omitted when micrometastases are found in sentinel lymph nodes (SLN). Although the 10-year follow-up results of IBCSG 23–01 [
8] and the American College of Surgeons Oncology Group(ACOSOG) Z0011 trial [
9] both support the avoidance of ALND in breast cancer patients with LNMM, most of these patients underwent breast-conserving surgery (BCS) plus whole breast radiation therapy (WBRT). Therefore, the conclusions of these two studies can only be applied to the clinical practice of relevant populations that meet the inclusion criteria and cannot be extrapolated to all pN1mi patients. In addition, the AMAROS trial [
10] showed that axillary radiotherapy (ART) is the best alternative to ALND in patients with 1–2 sentinel lymph node (SLN) metastases.
Accordingly, National Comprehensive Cancer Network(NCCN) guideline [
12] and American Society of Clinical Oncology(ASCO) guideline [
13] recommend radiotherapy of the axilla in lieu of ALND in patients with pathological SLN-positive and low tumor burden. Patients with LNMM are also treated in the light of the guidelines’ recommendation for patients with positive lymph nodes. However, different numbers of involved lymph nodes are all divided into the same N1mi stage without categorization. We designed this study to compare the prognosis and local treatment recommendations of N1mi breast cancer patients with different numbers of micrometastatic lymph nodes.
Patients and methods
Data source and study population
We screened the Surveillance, Epidemiology, and End Results (SEER) database of the National Cancer Institute to identify eligible breast cancer patients in this retrospective study. The SEER database is an open-access resource for cancer-based epidemiology and survival analyses (See Website “
https://seer.cancer.gov/data/” for detailed information). Data access for present study was authorized by SEER Program. As all patient information in the SEER database is de-identified, this study was exempt from Institutional Review Board evaluation.
The SEER*Stat version 8.4.0 was utilized to extract 32,032 pN1mi breast cancer patients’ information, diagnosed between January 2004 and December 2019 (Nov 2021 Submission). We excluded patients identified by death certificate or autopsy and with incomplete survival data. Female patients with T1-2 invasive breast cancer without distant metastasis and underwent breast surgery were included in this study. Patients who did not undergo surgery or whose type of surgery was unknown, had distant metastases, or had an unclear number of axillary lymph nodes examined were excluded (Supplemental Figure S
1). The data elements include patient basic demographic characteristics, cancer pathological types, staging and molecular biomarkers status, the treatment received for the cancer, and survival outcomes information. It is worth noting that the information of epidermal growth factor receptor-2 (HER2) status in the SEER database has been registered from 2010, so the correlation analysis of HER2 status and molecular types only included case data since then.
Finally, a total of 27,032 female breast cancer patients, with a stage of T1-2N1miM0, were included in our retrospective study. Patients were divided into three groups for prognosis comparison according to the number of LNMM: N1mi with 1 (Nmi = 1), 2 (Nmi = 2), or more (Nmi ≥ 3) involved lymph nodes. Furthermore, survival outcomes of different local treatments, including axillary surgery types and radiotherapy, are required for these three groups. The number of lymph nodes removed was used as a surrogate for the type of axillary surgery which was defined as in previous similar studies [
14‐
17], that is, patients with 5 or less lymph nodes resected were categorized as receiving sentinel lymph node biopsy (SLNB) while 6 or more as undergoing ALND.
Statistical analysis
Patients-, tumor-, and treatment- level characteristics are presented as frequencies (N) and percentages, and compared using chi-square tests as appropriate. Kaplan–Meier method and log-rank test were used to draw survival curves and compare differences among different subgroups without adjustment for other factors. Univariate and multivariate cox regression analyses were used to identify independent prognostic factors and to calculate hazard ratios (HR) toward target subgroups after adjustment for other prognostic factors. Stratified analyses and interaction analyses were also applied to explore the predictive significance of different LNMM involved numbers.
In the analysis of axillary surgery and radiotherapy, we adopted the method of 1:1 nearest propensity score matching (PSM) with matching tolerance 0.02, in order to balance the characteristic differences between the two compared axillary surgery groups, covariables included in propensity score matching were age, race, marital status, grade, T stage, nodal status, estrogen receptor (ER) status, progesterone receptor (PR) status, HER2 status, type of breast surgery, radiation and chemotherapy. All tests were two-sided, and a P value < 0.05 was considered to be statistically significant. All statistical analyses were performed using IBM SPSS software version 24.0 (IBM Corp., Armonk, USA) and R version 4.1.3 (The R Project for Statistical Computing, Vienna, Austria).
Discussion
From the 6th edition of AJCC manual [
2] for staging to the latest 8th edition [
14], no matter how many number of micrometastasis lymph nodes were detected, they were all divided into the same pN1mi staging, and the difference in prognosis caused by the inconsistent number was not distinguished. Evidence that the number of macrometastatic lymph nodes negatively affects survival outcome [
18‐
20] prompted refinement of the staging system. We designed this study to distinguish the prognosis and local treatment recommendations of N1mi breast cancer patients with different numbers of micrometastatic lymph nodes involved. Our study demonstrated that for breast cancer patients with identical T1-2N1miM0 stage, the greater number of LNMM, the worse the prognosis (
P < 0.001).
To investigate the prognostic significance of axillary lymph node micrometastases, the MIRROR trial [
21] was the first retrospective cohort study of patients with LNMM and ITC, which confirmed that for patients who did not receive adjuvant therapy, both pN0(i +) and pN1mi stages are independent prognostic indicators. There is no significant survival difference in prognosis between pN0(i +) and pN1mi stage patients, and both pN0(i +) and pN1mi stage patients can benefit from adjuvant therapy. The NSABP B-32 trial [
11,
22] enrolled 3795 breast cancer patients who underwent BCS and received postoperative whole-breast radiotherapy(WBRT) and systemic adjuvant therapy. After a median follow-up of 95 months, there was no significant difference in disease-free survival (DFS), OS and distant metastasis-free survival between pN1mi and pN0 stage patients. There are also discussions on the prognostic significance of the involved lymph node number in patients with pNmi stage. Roi Weiser et al. [
15] used National Cancer Database (NCDB) data to analyze the prognosis of lymph nodes status, and concluded all nodal status had a positive effect on survival compared with Nmic > 1 status, with HRs of 0.68, 0.88, and 0.93 for N0, Nmi = 1, and N1.1 disease respectively, with only N0 reaching statistical significance.
Through the SEER database registration data, it can be observed that before 2010, more pNmi patients chose to receive ALND, but this situation changed in 2011, and SLNB only became a preferred option for more patients, and the proportion of undergoing SLNB has since increased year by year. Until the most recent follow-up in 2019, about 79.84% of patients only received SLNB, and the proportion of patients with only one LNMM was as high as 82.14%. The management of the axilla in patients with LNMM has a long history. In 2010, Yi et al. [
23] reported a retrospective study on the choice of SLNB or ALND in patients with axillary lymph node-positive breast cancer. There were 6838 breast cancer patients with LNMM, of which 2240 received SLNB and 4598 received ALND, and post-surgery relevant systemic therapy and local radiation therapy. After 50 months of follow-up, there was no significant difference in the recurrence rate between patients who underwent SLNB only and those who underwent ALND. However, in our study, ALND does provide a significant survival benefit for N1mi breast cancer patients after a median follow-up of 95 months, whether in multivariate-adjusted cox regression analyses or survival analyses after PSM.
Both IBCSG 23–01 [
8] and ACOSOG Z0011 [
9] suggest that axillary dissection can be avoided in patients with early breast cancer and limited sentinel lymph node involvement. However, since more than 90% of the patients received BCS + whole breast radiotherapy in these two trials, the results are only applicable to the status of the enrolled population. AATRM trial [
24] is a prospective and randomized clinical trial specifically targeting the early breast cancer patients with sentinel lymph node micrometastases, it randomized patients to ALND or clinical follow-up and showed no significant difference in DFS between two groups. Another multi-institutional prospective study of 260 pT1-2Nmi post-mastectomy patients reported from Lim SZ et al. [
25] suggested that no statistically significant differences were found between patients with SLNB, ALND, or PMRT. Nonetheless, our study found discrepant results with different numbers of LNMM involved. SLNB with radiotherapy had the best prognosis when there was only one lymph node micrometastasis, while when the number of micrometastatic lymph nodes increased to two, ALND plus radiotherapy had the best survival outcome. Although only whether received radiotherapy or not achieved significant benefit, it still suggested that when the number of lymph node involved is different, the treatment mode should be focused and cannot be static.
The recently published prospective SENOMIC trial [
26] omitting a completion ALND in breast cancer patients with sentinel LNMM, and found that patients who had mastectomy without adjuvant radiotherapy had a significantly higher risk of recurrence than those who underwent breast-conserving surgery. As with trials such as IBCSG23-01 and Z0011, radiotherapy after BCS plays an important role in improving outcomes. In our stratification analysis of different treatments, we found that there was no significant difference in survival between different axillary surgery types, with or without radiotherapy. While when patients have undergone different types of axillary surgery, whether they receive radiotherapy become important. When pNmi breast cancer patients only underwent SLNB, compared with those who received radiotherapy, the HR value of without radiotherapy was 1.695 (95%CI 1.534–1.874;
P < 0.001).
Stratified analyses and interaction analysis in our study indicate that the benefit from local radiotherapy in pN1mi patients may be of even greater importance on the survival outcome. AMAROS trial [
10] testified axillary radiotherapy is the best option to replace ALND when 1–2 SLNs have metastasized in T1-2 breast cancer patients, which can improve the quality of life without affecting DFS and OS. The OTOASOR trial [
27] also have proved the equivalence of ALND and ART in patients with low lymph nodal burden. These two trials included 29% and 25% of patients with microscopic nodal disease respectively. In 2018, Wu SP et al. [
28] reported a retrospective study that evaluated the survival impact of PMRT in patients with N1mi within the National Cancer Database, and found that no OS differences were associated with PMRT, whether in the SLNB group or the ALND group. Another two large, single-institution studies separately from Memorial Sloan Kettering Cancer Center and MD Anderson Cancer Center [
29,
30] demonstrated no difference in local recurrence rates(LRR) for patients with N1mi disease post-mastectomy and SLNB, regardless of further radiation or ALND. However, a study from Merfeld EC et al. [
31] indicated that pN1mi patients with grade 3 were observed to be at substantial risk for LRR, and radiotherapy was associated with a lower risk of LRR.
Inevitably, there are several limitations related to its design and data source in our study. Firstly, this is a retrospective study derived from a public database, although PSM-based analyses can reduce the effects of the observed confounders, it cannot address unobserved confounders nor the inevitable cases-loses. Secondly, locoregional recurrence cannot be captured in the SEER database. And it is also unfortunate that cases receiving neoadjuvant chemotherapy (NAC) could not be identified in the SEER database, so axillary management cannot be discussed in patients receiving NAC.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.