Prognostic significance of occult lymph node metastases in breast cancer: a meta-analysis

Occult metastases in axillary lymph nodes have been reported to be associated with poor prognosis in patients with breast cancer. However, studies on the prognostic value of occult metastases have shown controversial results. This meta-analysis aimed to evaluate the prognostic significance of occult lymph node metastases in breast cancer. Studies published until May, 2020, which retrospectively examined negative lymph nodes by stepsectioning and/or immunohistochemistry, were retrieved from MEDLINE, EMBASE, CNKI, and Cochrane Library databases. The pooled Relative Risk (RR) with 95% confidence interval (95% CI) for overall survival (OS) and disease-free survival (DFS) were calculated to examine the associations between occult metastases and prognosis. Patients with occult metastases in axillary lymph nodes had poorer five-year DFS (RR = 0.930; 95% CI = 0.907–0.954) and OS (RR = 0.972; 95% CI = 0.954–0.990). Furthermore, the DFS (RR = 0.887; 95% CI = 0.810–0.972) and OS (RR = 0.896; 95% CI = 0.856–0.939) of patients with occult metastases were significantly lower after a ten-year follow-up. Occult metastases in the axillary lymph nodes are associated with poorer DFS andOS of patients with breast cancer. Occult metastases might serve as a predictive factor of survival outcomes in patients with breast cancer.


Background
Axillary lymph node (ALN) status is an important prognostic indicator of survival in breast cancer [1]. In 1948, Saphir et al. showed that a limited number of sections from the axillary lymph nodes of patients with breast cancer are insufficient to detect metastases [2]. Since then, occult metastases have been defined as metastases that were not initially assessed, but detected by further examinations [3]. Over the following decades, multiple new techniques have been introduced to improve lymph node biopsy. Using step-sectioning and immunohistochemical (IHC) staining, occult metastases have been frequently detected in 12-23% of women with breast cancer, who initially exhibit negative axillary lymph nodes on hematoxylin and eosin (H&E) staining during routine pathological examination [4][5][6][7][8][9][10]. In addition, some studies also used reverse transcriptase-polymerase chain reaction (RT-PCR) to detect specific mRNA [6,7,11].
The prognostic significance of occult metastases remains controversial. Although several studies indicated occult metastases impacts OS or DFS [9,[12][13][14][15], others argued that occult metastases have no significant prognostic value [6,10,16,17]. Furthermore, the routine use of IHC to stage lymph nodes has beenquestioned in recent large sample size studies. The National Surgical Adjuvant Breast and Bowel Project randomized controlled trial B-32 (NSABP B-32) indicated that occult metastases were an independent prognostic variable in survival, however, the difference in outcome at 5-years was small (1.2 percentage points) [14]. The American College of Surgeons Oncology Group (ACOSOG) Z0010 study also demonstrated that IHC evidence of occult metastases was not significantly associated with OS [18]. Hence, the current National Comprehensive Cancer Network (NCCN) guideline for breast cancer does not recommend routine IHC to define node involvement [19]. Although several systematic reviews have been published on the association between occult metastases and survival [3,16], an update including recent studies is still necessary. This meta-analysis systematically evaluated the association between occult lymph node metastases and survival among patients with breast cancer.

Methods
The present systematic review and meta-analysis was performed according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.

Search strategy
The literature review was performed in PUBMED, EMBASE, China National Knowledge Infrastructure (CNKI), and Cochrane Library until May 1, 2020. The following search terms were used: (breast cancer OR breast carcinoma OR breast neoplasms), (lymph node OR lymph-node), (occult metastases OR micrometastases OR isolated tumor cells), and (prognosis OR prognostic OR survival OR survival rate OR survival analysis OR mortality OR recurrence). Relevant reviews, metaanalyses, and references cited in these papers were also checked for potential studies. Abstracts or unpublished reports were not considered. If more than one article was published by the same author using the same case series, the study with the highest number of subjects was selected. All the searches were conducted by two reviewers independently, and any disagreement was resolved through discussion.

Inclusion and exclusion criteria
The following inclusion criteria were applied: ①evaluation of the association between occult metastases and survival outcome of breast cancer patients, ② casecontrol or cohort design, ③description of the survival outcomes of the studies in terms of disease-free or overall survival, and ④full texts based on original data. The exclusion criteria employed were: ①no control group (lymph node-negative group), ②lack of Kaplan-Meier methods or life-table analyses, and ③short follow-up period (< 5 years).

Data extraction and study quality assessment
All data were extracted independently by two reviewers, according to pre-specified selection criteria. Disagreement was resolved by consensus and discussion with the third investigator. The following data were extracted: pathological assessment of the removed lymph nodes, number of control group, number of patients with lymph node occult metastases, tumor stage, follow-up period, performance of axillary lymph node dissection, disease-free/overall survival rates, administration of adjuvant systemic therapy, and results of multivariable analyses (Table 1). If the survival data were not provided in a table or text in the chosen articles, they would be extracted from the survival curves by Engauge Digitizer version 10.8 (GitHub, Open Source software). To prevent overlap of the data from studies that described subpopulations besides a total population, only subpopulations were taken into account for the calculation of relative risks. The Newcastle-Ottawa scale (NOS) was applied to assess the quality of the study [38].

Statistical analysis
The five-and ten-year relative risk (RR) of disease-free and overall survival was compared between the occult metastases group and control (lymph node negative) group. Statistical heterogeneity was measured using I 2 (I 2 > 50% was considered statistically significant heterogeneity). If significant heterogeneity was detected, the random-effects model was used and sensitivity analysis performed by removing one study at a time to calculate overall homogeneity and effect size. Otherwise, a fixedeffects model was employed [39]. Egger's regression method was used statistically to assess publication bias (p < 0.05 was considered statistically significant) [40]. Statistical analysis was performed using STATA version 12.0 (StataCorp LLC, US), 95% confidence intervals (CIs) were reported, and p < 0.05 was considered statistically significant. All statistical tests were two-sided.

Evaluation of publication bias
As including almost all studies, the studies of 5-yr DFS were conducted by using funnel plots and Egger's test to assess publication bias. The funnel plot was approximately symmetrical (Fig. 7) and the result of Egger's test (P = 0.567) revealed no obvious publication bias among the studies.

Discussion
This meta-analysis of 36 studies aimed to explore the association between occult metastases in axillary lymph nodes and long-term prognosis ofpatients with breast cancer. In this meta-analysis, occult metastases were associated with worse DFSand OS than nagetive lymph nodes after 5-years and 10-years' follow-up. These results demonstrated that occult metastases could be an independent prognostic factor in breast cancer patients with negative nodes on initial biopsy. Although the result of the sensitivity analysis were consistent, occult metastasis showed a worse prognosis after excluding the NSABP B-32 study [14].. This heterogeneity might be due to the second largest sample size and the relatively smaller difference in 5-year survival. Although the included studies differed in terms of patient population, pathological assessment, follow-up duration, and methodology, their results were generally homogenous. This is not the first meta-analysis to evaluate the relationship between occult metastases and survival in patients with breast cancer, howevwe, it has several strengths over the previous meta-analysis [3]. The previous meta-analysis pointed out that presence of occult metastases was associated with poorer DFS and OS. This current metaanalysis included recent large sample size studies and obtained consistent results [14,38]. The results were predominantly consistent with two previous large sample size studies, which albeit demonstrated that patients with occult micrometastases in axillary lymph nodes had a poorer survival [14,37]. Weaver et al. [14] found that occult metastases were an independent prognostic factor in patients with sentinel lymph nodes (SLN) that were negative on initial examination. However, they indicated that additional evaluation, including IHC analysis, had no clinical benefit because the difference between survival was statistically significant relatively small. A larger study including 93,070 patients also demonstrated a difference in OS between patients with occult metastases and those with IHCnegative lymph nodes [37]. However, in further multivariate analysis in subgroups, micrometastases (0.2-2 mm diameter), rather than isolated tumor cells (< 0.2 mm diameter or < 200 cells, ITC), remained an independent predictor for survival.
The results of this meta-analysis differ from those of ACOSOG Z0010 study, which demonstrated no obvious difference in 5-year OS between occult metastases and no metastasis [18]. The difference might be due to that ACOSOG Z0010 was limited to early-stage T1 and T2 tumors. Several studies included in our analysis also reported that occult metastases could not predict a poorer survival in patients with breast cancer [6,11,20,21,32,35]. This lack of significant difference might be due to the small sample size and the pathological examination techniques. It is worth mentioning that the Z0011 study has led to significant changes in the surgical treatment of breast cancer [41]. Fewer patients now receive further axillary surgery even in the presence of macrometassis. However, similar to the Z0010 study, the patients enrolled in the Z0011 study were also relatively at a lower risk: the tumor stage was T1-2, of which about 70% were T1 and 71% had only one positive SLN. Moreover, all patients underewnt breast-conservation surgery, and more than 96% of them received subsequent adjuvant systemic therapy. Therefore, the impact of occult metastasis on the survival of patients with a higher risk of recurrence (larger tumor size, mastectomy without adjuvant treatment, etc) needs further study.
The size of lymph node metastasis and subsequent treatment may affect the survival of patients with occult metastases. Weaver et al. [14] reported that the hazard ratio for death was 1.38 (95% CI = 1.02-1.87) in patients with isolated tumor cells and 1.91 (95% CI = 1.41-2.59) in those with micrometastases or macrometastases, when compared with patients in whom occult metastases were not detected. The MIRROR study showed that ITCs or micrometastases in SLN were associated with a lower 5-year DFS among patients who did not receive adjuvant therapy. Meanwhile, DFS was improved in patients with ITCs or micrometastases who received adjuvant therapy [42]. Consistent with the above results, this meta-analysis found that occult metastasis in axillary The relationship between axillary lymph node ITC/ micrometastases and recurrence has not been fully established. Although studies have confirmed that adjuvant systemic therapy can improve the prognosis of patients, current evidence is still insufficient to support the replacement of axillary lymphadenectomy (ALND) with adjuvant systemic therapy alone. A further analysis of the MIRROR trial showed that patients with SLN micrometasis who did not undergo axillary treatment had an higher 5-year regional recurrence rate [43]. Therefore, we recommend to improve the detection rate of occult metastases in SLNs, especially for those patients who may be exempted from adjuvant therapy. In addition, adjuvant systemic therapy and axillary local radiotherapy should be performed in all patients with micrometastases, while ALND to minimize complications of surgery and improve quality of life while achieving a good prognosis.
The limitations of this study are as follows: (1) this meta-analysis was based on data from survival curves instead of pooled individual data; (2) pathological type, surgical options, adjuvant treatment regimen, and systemic therapy may be associated with DFS and OS, but these detailed data were not available in the majority of included studies or stratified analysis could not be conducted; (3) the majority of included studies had a retrospective design, so recall and selection biases may affect the results; (4) publication bias is an inevitable problem since this study is based on published articles, and ongoing or unpublished studies were not included in this meta-analysis.

Conclusions
In summary, we found that occult metastases in the axillary lymph nodes of patients with breast cancer are an independent predictor of disease-free and overall survival. Moreover, it may indicate a relatively poor prognosis. However, because of non-standardized pathological examination and treatment, the prognostic value of occult metastases is still limited and further study is needed.