Proposed Modification of the pN2 Classification of the 8th Edition AJCC Staging System for Esophageal Squamous Cell Carcinoma: A Preliminary Study Based on the Chinese Population

Objective To evaluate the efficacy of the nodal staging of the 8th edition AJCC staging system for esophageal squamous cell carcinoma (ESCC) and propose a modification of the current pN2 classification. Methods 1188 patients who underwent esophagectomy for ESCC at Sun Yat-sen University Cancer Center in Guangzhou (Guangdong, China) between January 2005 and June 2010 were reviewed. We used the X-tile software to determine the optimal cutoff points. Kaplan–Meier method and log-rank test were used to compare the differences of survival. Multivariate Cox regression analysis was performed for the factors that were statistically significant in univariate analysis. Result In multivariate Cox regression analysis, alcohol consumption, pT status, and pN status were independent prognostic factors for overall survival (OS) according to the current pN classifications. And the observed 5-year OS rates for groups pN0, pN1, pN2, pN3 were 66.7%, 45.0%, 31.5%, and 21.5%, respectively (P<0.001). Based on the above results, the current pN2 classification was further subdivided as pN2a [3 metastatic lymph nodes (LNs)] and pN2b (4−6 metastatic LNs) groups. The 5-year OS rates for groups pN0, pN1, pN2a, pN2b, and pN3 were 66.7%, 45.0%, 37.7%, 26.3% and 21.5%, respectively (P<0.001). The rate of 5-year disease-free survival (DFS) was 60.0% for patients with pN0, compared with 36.8%, 29.3%, 20.8%, and 14.3% for those with pN1, pN2a, pN2b, and pN3, respectively (P<0.001).The current pN2 classification should be subdivided as pN2a (3 metastatic LNs) and pN2b (4–6 metastatic LNs) groups. The modified pN2 classification could better discriminate the survival differences between patients with 3–6 metastatic LNs for ESCC in the Chinese population.


Introduction
Globally, esophagealcancer (EC) is the seventh most common diagnosed cancer and the sixth most frequent cause of cancer-related deaths [1,2]. e histological type of EC mainly consisted of squamous cellcarcinomaand adenocarcinoma [3]. In China, more than 90% of EC cases is squamous cell carcinoma [4,5]. e prognosis of EC is still unsatisfactory and the 5-year overall survival (OS) rate is not more than 25% [6,7].
Lymph node (LN) status is one of the most important factors for the prognosis of EC patients [8]. Tumor-nodemetastasis (TNM) staging system is applied to estimate prognosis and guide treatment plan making in clinical practice [9]. e current nodal staging system is based on the number of metastatic LNs. According to the eighth edition of the American Joint Committee on Cancer (AJCC) staging manual, the pN categories of EC are classified as pN0 (0), pN1 (1-2), pN2 (3)(4)(5)(6), and pN3 (≥7) based on the number of metastatic LNs [10]. e current nodal staging system defines patients with 3-6 metastatic LNs as pN2, which is too general. e previous studies have demonstrated that the number of metastatic LNs is significantly associated with prognosis of EC [11][12][13].
us, the current pN2 classification should be divided into more subgroups.
In this study, data were collected from a large cohort of Chinese patients in a high-volume institution. We aimed to provide a proposal to subdivide the current pN2 classification, which could provide more prognostic information for esophageal squamous cellcarcinoma (ESCC) patients in the Chinese population.

Patients Selection.
Patients who underwent esophagectomy for ESCC at Sun Yat-sen University Cancer Center in Guangzhou (Guangdong, China) between January 2005 and June 2010 were reviewed. is study was approved by the Ethics Committee of Sun Yat-sen University Cancer Center.
Patients were included based on the following eligibility criterion: patients who underwent radical esophagectomy with pathologically confirmed ESCC. Patients were not eligible according to the following criteria: (1) diagnosed with not squamous cell carcinoma; (2) patients with preoperative therapy, including radiotherapy, chemotherapy, or chemoradiotherapy; (3) patients who underwent R1/R2 resection; (4) patients with a second tumor; (5) those who died within 30 days of surgery; (6) patients with carcinoma in situ; (7) patients with incomplete clinicopathologic information; and (8) patients with distant metastasis. Finally, 1188 patients were included in the present study ( Figure 1).

Follow-Up.
All patients were followed up after surgery every 3 months for the first year, every 6 months for the next 2 years, and annually thereafter. e follow-up examinations including thoracoabdominal CT, endoscopy, tumor markers, and barium esophagography. e last follow-up date was conducted at July 1, 2015. e interval between the date of surgery and the date of death or the last follow-up was defined as overall survival (OS) time. e disease-free survival (DFS) was defined as the time period from surgery to the appearance of tumor progression or the date of cancer-related death.

Statistical Analysis.
Statistical analyses were performed using SPSS 25.0 software (SPSS Inc., Chicago, IL, USA). e optimal cutoff points were identified using the X-tile software (version 3.6.1, Copyright Yale University, 2003) [14]. X-tile software determined the best cutoff points by using the minimum p value. Kaplan-Meier method and log-rank test were used to compare the differences of survival. Multivariate Cox regression analysis was performed for the factors that were statistically significant in univariate analysis. A two-tailed P value <0.05 was considered statistically significant.
Cox univariate and multivariate analyses were used to evaluate the prognostic factors of OS based on the current pN classifications. Univariate analysis demonstrated that smoking status, alcohol consumption, tumor length, pT status, pN status, and differentiation were independent prognostic factors of OS. In multivariate analysis, only alcohol consumption, tumor length, pT status, and pN status were independent prognostic factors ( Table 1). Kaplan-Meier curves were applied to compare the survival rates among different pN groups. e analysis results demonstrated that the observed 5-year OS rates for groups pN0, pN1, pN2, and pN3 were 66.7%, 45.0%, 31.5%, and 21.5%, respectively (P<0.001, Figure 2). e 5-year OS rate was significantly worse with the increasing pN classifications.
Based on the above results, the current pN2 classification was further subdivided as pN2a (3 metastatic LNs) and pN2b (4-6 metastatic LNs) groups by using the X-tile software ( Figure 3). Based on the revised pN classifications, univariate and multivariate analyses were performed to identify the risk factors of OS and DFS. In univariate analysis, smoking status, alcohol consumption, tumor length, pTstatus, revised pN status, and differentiation were associated with OS and DFS significantly. However, only alcohol consumption, tumor length, pT status, and revised pN status were independent prognostic factors of OS in multivariate analysis (Table 2). Moreover, only tumor length, pT status, and revised pN status were significantly associated with DFS in multivariate analysis (Table 3). e 5-year OS and DFS rates based on the revised pN classifications were shown in Figures 4(a) and 4(b). e 5-year OS rates for groups pN0, pN1, pN2a, pN2b, and pN3 were 66.7%, 45.0%, 37.7%, 26.3%, and 21.5%, respectively (P<0.001). e rate of 5-year DFS was 60.0% for patients with pN0, compared with 36.8%, 29.3%, 20.8%, and 14.3% for those with pN1, pN2a, pN2b, and pN3, respectively (P<0.001). Increased pN classifications demonstrated significantly decreased survival.

Discussion
LN status is an important prognostic factor for EC patients. In clinical practice, the 8th edition AJCC classification system is widely applied in staging of EC. e 8th nodal staging defines 3-6 metastatic LNs as pN2 in EC, which is too general and could not provide more detail of the survival information.
In the present study, we validated the current pN classifications of the 8th edition AJCC classification system in the first stage and found that there were significant differences in survival between different pN    groups. After that, the current pN2 classification was subdivided as pN2a and pN2b groups. What was more, patients with pN2a had a better survival than those with pN2b significantly and significant survival differences existed between each revised pN classification. e results of our study showed that the revised pN2 classifications could provide more accurate prognosis information for EC patients. As a reasonable nodal classification system, there should be a higher predictive value for prognosis and a different Journal of Oncology 5 survival rate between each group. Although the current nodal staging provided a more exact evaluation of survival than the sixth edition, it had not the ability to distinguish the survival of EC patients with 3-6 metastatic LNs [15]. However, the revised pN2 classification showed a good discriminatory ability for survival in these patients. e current pN2 classification of EC did not consider the effect of the specified number of metastatic LNs, which might not be perfect.
Previous studies demonstrated that a nodal staging system based on the number of metastatic LNs may be more sensitive for the survival estimation for EC than  Journal of Oncology classification based on the location of LNs [16][17][18]. Compared with other indicators such as LN ratio (LNR, the ratio between the number of metastatic LNs and the total number of examined LNs) and log odds ratio (LODDS, the log of the ratio between the number of positive LNs and the number of negative LNs), the indicator-the number of metastatic LNs-was simpler and thus easy to generalize. us, our proposal to subdivide the pN2 classification was still based on the number of metastatic LNs. In univariate and multivariate analyses, we also found that alcohol consumption was an independent prognostic factor of survival.
is was consistent with previous studies [19,20]. However, no significant survival differences were observed between the patients with <15 LNs examined and those with ≥15 LNs examined (P�0.073). is might be because of the sample size of this study and the bias of retrospective study. e P value (0.073) was closed to 0.05 and we thought that, with the sample size increased, there would be significant survival differences between these two groups.
It is well known that the clinicopathological features of EC vary widely between patients in Asian and Western countries [21]. Squamous cell carcinoma is the most common pathological type in China, and it is considered that more data from Chinese patients are essential to validate the generalizability of the 8th edition AJCC nodal staging system for ESCC [5]. A relatively large sample size solely from Chinese ESCC patients in one institution was included in this study, which could help to decrease the heterogeneous characteristics. Additionally, we determined the optimal cutoff values using the X-tile, which was one of the most convenient and objective methods [14,22]. As such, the results from our data were more precise and significant, which could better represent the prognostic characteristics of ESCC in the Chinese population.
Some limitations in the present study should be acknowledged. Firstly, the bias of a retrospective investigation was unavoidable. Secondly, the number of pN3 cases was small, which implied that the results needed further validation. is was a preliminary study based on single-center and we planned to enroll other databases to verify the conclusion and further subgroup the pN2 classification next.  In conclusion, we proposed a modification of the current pN2 classification for ESCC. e modified pN2 classification could better discriminate the survival differences between patients with 3-6 metastatic LNs for ESCC in the Chinese population.

Data Availability
We may balance the potential benefits and risks for each request and then provide the data that could be shared.

Conflicts of Interest
e authors declare that they have no conflicts of interest.