Metastatic Axillary Lymph Node Ratio ( LNR ) is Prognostically Superior to pN Staging in Patients with Breast Cancer-- Results for 804 Chinese Patients from a Single Institution

Breast cancer is the most commonly diagnosed cancer among women in China and numerous other parts of the world. Axillary lymph nodes (ALN) status is one of the most important prognostic factors in breast cancer and the number of involved nodes has been incorporated into routine clinical decision making (Fitzgibbons et al., 2000). According to guidelines of the 7th edition American Joint Committee on Cancer (AJCC) staging system, patients with different positive ALNS have been classified as various pN disease (pN0: zero positive node, pN1: 1–3 positive nodes, pN2: 4-9 positive nodes and pN3: ≥ 10 positive nodes) (Singletary et al., 2002; Vinh-Hung et al., 2003). Recently, however, a few studies have suggested that, not only the number of postive nodes but also the total number of lymph nodes removed should be taken into consideration. Therefore, a new prognostic factor of lymph node ratio (LNR), which was defined as the ratio of the number of positive to total removed lymph nodes,


Introduction
Breast cancer is the most commonly diagnosed cancer among women in China and numerous other parts of the world.Axillary lymph nodes (ALN) status is one of the most important prognostic factors in breast cancer and the number of involved nodes has been incorporated into routine clinical decision making (Fitzgibbons et al., 2000).According to guidelines of the 7th edition American Joint Committee on Cancer (AJCC) staging system, patients with different positive ALNS have been classified as various pN disease (pN0: zero positive node, pN1: 1-3 positive nodes, pN2: 4-9 positive nodes and pN3: ≥ 10 positive nodes) (Singletary et al., 2002;Vinh-Hung et al., 2003).Recently, however, a few studies have suggested that, not only the number of postive nodes but also the total number of lymph nodes removed should be taken into consideration.Therefore, a new prognostic factor of lymph node ratio (LNR), which was defined as the ratio of the number of positive to total removed lymph nodes,
In many of the above studies, however, additional factors affecting DFS, such as tumor receptor status, HER2 expression, and use of neoadjuvant or adjuvant treatment regimens, were not considered.The cutoff values were often various in different institutions as well.What's more, up to now, the recommendation of using LNR as an alternative to the current pN staging has no fomal guideline to follow.Hence, in this article, we sought to determine the prognostic impact of LNR in comparison to pN staging and other established prognostic factors in Chinese breast cancer patients for the first time.Based on the caculated optimal cutoff values, we then further evaluated the independent prognostic value of LNR compared to pN staging in predicting DFS of breast cancer.

Materials and Methods
This study population consisted of 804 female patients who were confirmed to have the diagnosis of nonmetastatic primary invasive breast cancer and have underwent axillary lymph nodes dissection at Sun Yat-Sen University Cancer Center from 1999 to 2008.After institutional review board approval, we used medical records and tumor registry information to collect data on these patients.Patients were excluded from this study who (1) had distant metastases at diagnosis; (2) had incomplete pathological data entries; (3) underwent neoadjuvant treatment; (4) were lost during follw-up; and (5) had secondary malignancies except for non-melanoma skin cancer or in situ cervix cancer.Patients received treatment according to the standard care.Patient-related characteristics (age, body mass index, pT stage, pN stage, LNR, pTNM stage, estrogen/progesterone receptor status, and human epidermal growth factor receptor 2 status), and treatment-related factors (chemotherapy, radiotherapy and hormone therapy) were analyzed.
Postoperative follow-up included physical and laboratory examinations every 3 months for the first 3 years, every 6 months for the fourth and fifth years, and annually after 5 years.End point for survival analyses was disease-free survival (DFS), which was defined as time from operation to documentation of the first of the following events: any recurrence (local, regional, or distant) of breast cancer, a contralateral breast cancer, a second primary cancer, or death due to any cause.Patients who were alive without any of these events were censored at the date of their last disease evaluation.Follow-up was provided for the entire study population until July 2011.

Statistical analysis
Comparisons were made by chi-square analysis for categorical variables and 2-sided t tests for continuous variables.A Cox proportional hazards model was used to compare parameters and to adjust for known prognostic clinical and pathological variables.DFS was calculated using the Kaplan-Meier method and the differences in survival time were estimated by the log-rank test.A multivariate Cox regression analysis (forward stepwise) was carried out to compare and identify independent prognostic factors.The 95% confidences interval (95% CI) was used to quantify the relationship between survival time and each independent factor.Differences were considered to be significant if the P values from a

Results
There were 804 cases with breast cancer identified for analysis and the corresponding clinicopathological characteristics were shown in Table 1.The 5-year diseasefree survival rate for patients with pN positive or negative were 96.6% and 80.0%, respectively, and 713 patients were alive when our follow-up was completed (Figure 1).

Univariate and multivariate survival analyses
Seven factors were found to have statistically significant associations with the disease-free survival on univariate analysis, which were pathological T (pT), pN, LNR, HER2 status, Molecular type, pTNM stage and Radiotherapy (Table 1).All the aforementioned 7 variables were included in a multivariate Cox proportional hazards model to adjust for the effects of covariates.Based on the number of pN stage, 196 patients are classified as pN1, 101 as pN2 and 64 as pN3.Using the LNR stage, 223 patients are classified as LNR1, 98 as LNR2 and 40 as LNR3.The 5-year survival rates according to each pN and LNR stage are shown in Figure 2.With both the pN and the LNR stage, the Kaplan-Meier plots showed a good discriminatory ability (Figure 2A and 2B).The 5-year disease-free survival(DFS) rates of LNR0, LNR1, LNR2, and LNR3 patients were 97.1%, 92.4%, 68.9%, and 52.0%, respectively (log-rank x 2 = 164.879).In addition, pN classification showed an imbalance in prognostic separation, with the pN2 and pN3 survival curves closed to each other, while the LNR curves remained separated exceeding 5 years.For patients in the pN2 category, the five year DFS rates for LNR1-3 were 88.1%, 69.1%, and 25.0%, respectively (P<0.05;Table 2 and Figure 3).However, when we compared the survival rates among the pN categories in different LNR categories, there was no significant prognostic difference between patients in different pN categories for any any LNR categories.By comparison, the 5 years DFS rates for patients with pN0, pN1, pN2, and pN3 stage were 97.1%, 92.5%, 70.2%, and 65.9%, respectively (log-rank x 2 = 134.792).Compared with patients with pN1 stage, the adjusted hazard ratio of DFS was 4.211 for patients with pN2 and 6.324 for patients with pN3, while 3.921 for patients with LNR2 and 6.324 for patients with LNR3 compared with patients with LNR1 stage.
In multivariate survival analyses, we demonstrated that only LNR remained independent prognostic factors (data not shown).

Discussion
Although the current TNM classification system is still the mainstay of breast cancer staging, the role of the total number of lymph nodes retrieved is still unknown (Li et al., 2012;Wang et al., 2012).To ensure optimal treatment for individual breast cancer patients in this era of personalized medicine, it is necessary to escalate the use of AJCC staging system.Recently, the concept of LNR has aroused deep concern around the world and the role of pN has been challenged (Kim et al., 2011).The positive nodes harvested in the axilla are dependent on the surgeon (technique and philosophy), the pathologist (specimen evaluation technique and philosophy), and the patient (tumor site, T stage, immune response, and age) (Chen et al., 2011).So, any of the above factors could affect the pN staging, and then patients may be inaccurately staged.
However, the LNR-namely, the ratio of the number of positive to total removed lymph nodes-attempts to standardize against the inconsistency and variability of nodal assessment.Previous studies have demonstrated that the effects of LNR staging is superior to that of pN staging in predicting prognosis of patients for many cancer sites: colon (Chen et al., 2011), gastric (Wang et al., 2011;Yu et al., 2011), pancreas (Pawlik et al., 2007;Keck et al., 2012), esophagus (Greenstein et al., 2008), corpus uteri (Chan et al., 2007) and bladder (Herr, 2003).
As expected, in our study, we demostrated that the power of LNR was much better than pN in predicting the breast cancer DFS.Then, for the first time in Chinese breast cancer patients, we proposed a categorization of the LNR which was validated by bootstrap resampling among patients.Patients in our study were then classified into four groups (LNR0, 0%; LNR1, 1%-30%; LNR2, 31%-80%; and LNR3, 81%-100%).
Our cutoff points that would be required for a staging classification are different from previous studies.Tsuchiya A et al. retrospectively reviewed the data for 37 breast cancer patients treated between 1987 and 1995.The corresponding cutoff points of LNR were 0.10 and 0.50 (Tsuchiya et al., 1997).Walker MJ et al. retrospectively reviewed the records of 141 patients who had breast cancer with > 10 positive nodes and they defined the cutoff points with 0.50 and 0.75 (Walker et al., 1995) Voordeckers M et al. also demostrated that the LNR was the most important factor in 810 breast cancer patients who are lymph nodepositive, the LNR were categorized into three groups by 0.10 and 0.50 (Voordeckers et al., 2004).Vinh-Hung et al. analyzed 1,827 breast cancer patients with all T and N stages and demostrated prognostic importance of LNR, and then categorized the LNR into different risk groups with 0.20 and 0.65 (Vinh-Hung et al., 2009).To identify cutoff points that should be stable for specific population, we chose the X-tile software method to calculate the cutoff points and validated them with bootstrap method.In our cohort, the caculated cutoff points from X-tile for LNR were compatible with our cancer center population.Therefore, in distrinct areas, the cutoffs should be modified when the LNR was the significant factor predicting prognosis for breast cancer patients.
Moreover, the LNR in our research could discriminate subsets of patients with similar prognosis in pN staging: the five years DFS of LNR 1-3 were 88.1%, 69.1%, and 25.0% for patients in the pN2 category, respectively (P<0.05).These findings showed the ability of LNR classification to neutralize the risk of stage migration.
However, we acknowledge several limitations in our study.Firstly, it was retrospective and single center, then it might lead to bias in patients seletion.Moreover, the overall survival time would be obtained in the future.Therefore, the independent validation of LNR as a prognostic factor in breast cancer awaits further prospective clinical trials.
In conclusions, the LNR staging identified three cutoff points which could define breast cancer prognosis more accurately than the pN staging.As the LNR staging is an available and inexpensive prognostic factor for patients with breast cancer, therefore, it should be incorporated into the current AJCC staging system in future.

Figure 1 .Figure 3 .
Figure 1.Kaplan-Meier DFS for Patients with Breast Cancer According to Lymph Node Status

Table 2 . DFS Rates on the Basis of pN Category According to the LNR Category
patients DFS rate (%) patients DFS rate (%) patients DFS rate (%)