Introduction

Esophageal cancer is the sixth leading cause of cancer deaths worldwide [1]. The overall survival of esophageal cancer remains poor although some improvements have been achieved in treatment [1, 2]. Lymph node status is one of the factors closely related with the long-term survival of esophageal cancer patients. Esophageal carcinoma patients with nodal metastases generally have a poor prognosis after surgical resection. Patients without involvement of regional nodes (pN0) had a significant improvement in 5-year survival [3].

The seventh American Joint Committee on Cancer (AJCC) tumor–node–metastasis (TNM) staging system includes primary tumor status, lymph node status, and metastatic status. N staging predicting esophageal cancer patient prognosis has been studied [4–7]. However, N stage is categorized by the number of positive lymph node retrieved, which can be influenced by many factors, such as variation in examining pathologists, the adequacy of the surgical resection, obesity, and so on [8–11].

The concept of lymph node ratio (LNR) was discussed by some reports recently. LNR, the ratio of metastatic to total retrieved nodes, which is considered to reflect metastatic lymph node status more accurately, has been found to have prognostic significance in several kinds of tumors [12–16]. However, in esophageal cancer, there are limited studies about the prognostic significance of LNR. Hence the aim of our study was to investigate whether LNR can be functioned as an independent prognostic factor of esophageal cancer.

Patients and methods

Patients

We collected all esophageal cancer patients who underwent esophagectomy at the Department of Thoracic Surgery, Qilu Hospital of Shandong University, from January 2007 to December 2008. Patients were excluded from the present study when they (1) died within 30 days after operation, (2) underwent preoperative chemotherapy or radiotherapy, and (3) could not be contacted during follow-ups. A total of 209 cases were included in this study. All patients included in our study had histologic documentation from at least two independent pathologists. Among 209 cases, 111 patients did not receive any therapy after operation, 45 cases received postoperative radiotherapy, 22 cases received postoperative chemotherapy, and 31 cases received postoperative chemoradiotherapy. Clinicopathological data was obtained from the hospital electronic recording system, such as gender, tumor location, tumor stage, lymph node metastasis, treatment regimen, and so on.

Lymph node ratio (LNR)

LNR was the ratio of metastatic lymph nodes to total retrieved nodes. Patients were divided into two groups according to LNR for analysis. To determine the optimal cutoff point for the number of the lymph node ratio as predictors of survival, we simulated the log-rank test for groups defined by LNR < c) and LNR > c for observed values of the covariate for the entire data set. The largest log-rank test statistic was then applied to detect the optimal cutoff point [17].

Follow-up

The follow-up was updated to October 2013. Survival time was calculated from the date of surgery to the event or the last follow-up. The reexamination included clinical examination, chest X-ray, abdominal CT scan, colonoscopy, and evaluation of CEA and CA19-9.

Statistical analysis

All statistical analyses were performed using the SPSS statistical software package (version 20.0; SPSS Inc, Chicago, Ill). Survival analyses were performed by Kaplan–Meier curves with log-rank tests for significance. Statistical analyses included univariate analysis and multivariate analysis. Univariable Cox regression analyses were performed using disease recurrence or death as the outcomes with a significance level of p < 0.05. Multivariate analysis was carried out with a Cox proportional hazards model to evaluate LNR and other prognostic factors with respect to disease-free survival (DFS) and overall survival (OS). Two-sided p values of <0.05 were considered statistically significant.

Results

Clinicopathological characteristics of patients

Among the 209 cases, 80.4 % was male and 19.6 % was female. The median age of this cohort was 60.6 years. Most of primary tumor sites were middle thorax (56.5 %) and lower thorax (33.5 %). Patients with a LNR less than 0.2 accounted for 76.6 % of the whole cohort. Table 1 shows the clinicopathological parameters of the studied patients.

Table 1 Patient characteristics
Full size table

Prognostic impact of LNR and N stage

The median DFS of this cohort was 35.2 months, and 5-year DFS rate was 32.1 %. The median OS of this cohort was 46.4 months, and 5-year OS rate was 40.0 %. Kaplan–Meier survival analysis revealed a correlation between LNR and overall and disease-free survival times. Patients with LNR higher than 0.2 had significantly poorer DFS (p < 0.001) and OS (p < 0.001) than those with LNR less than 0.2. Kaplan–Meier curves of DFS and OS based on LNR are shown in Figs. 1 and 2.

Fig. 1
figure 1

Kaplan–Meier curve of OS based on LNR

Full size image
Fig. 2
figure 2

Kaplan–Meier curve of DFS based on LNR

Full size image

Survival analysis showed significant difference in OS (p < 0.001) and DFS (p < 0.001) among different N categories in this entire data. Kaplan–Meier curves of DFS and OS based on N stage are shown in Figs. 3 and 4.

Fig. 3
figure 3

Kaplan–Meier curve of OS based on N stage

Full size image
Fig. 4
figure 4

Kaplan–Meier curve of DFS based on N stage

Full size image

Univariate and multivariate survival analyses

The results of univariate analyses are shown in Table 2. The log-rank test revealed that factors significantly correlated with OS and DFS, including T stage, N stage, LNR, TNM stage, histological grade, and treatment regimen (all p < 0.05).

Table 2 Univariate analysis of factors associated with OS and DFS
Full size table

Multivariate analyses were performed using Cox proportional-hazards regression. Table 3 shows the results of multivariate analyses. In multivariate analysis, LNR was found to be an independent prognostic factor for DFS (p = 0.008, HR 1.863, 95 % CI 1.180–2.942) as well as TNM stage (p < 0.001, HR 2.215, 95 % CI 1.424–3.444). With respect to OS, LNR was also an independent prognostic factor (p = 0.025, HR 1.708, 95 % CI 1.068–2.731). Besides, N stage (p = 0.028, HR 1.626, 95 % CI 1.055–2.506), TNM stage (p = 0.002, HR 2.051, 95 % CI 1.304–3.226), and treatment regimen (p = 0.047, HR 1.198, 95 % CI 1.002–1.432) were independent prognostic factors for OS.

Table 3 Multivariate analysis of factors associated with OS and DFS
Full size table

Subgroup analysis

Significant differences in OS and DFS rates were found between different LNR categories within the same N category (p < 0.05, Table 4) but not between different N categories within the same LNR category (p > 0.05, Table 5). These results confirmed that LNR showed better prognostic value than N stage for esophageal cancer.

Table 4 OS and DFS based on N category according to the LNR category
Full size table
Table 5 OS and DFS based on LNR category according to the N category
Full size table

Discussion

Esophageal cancer staging relies on the TNM system designed jointly by the Union International Against Cancer (UICC), and the AJCC. TNM staging system is the international unifying classification for malignancy, including primary tumor status (T), metastatic lymph node status (N), and distant metastatic status (M). The UICC and the AJCC have provided a more accurate estimation of prognosis than the sixth edition, proposing an updated classification (seventh edition) for N categories based on the number of metastatic lymph nodes [18], which emphasizes the significance of metastatic lymph node on the clinical staging. Thus, more and more attention has been focused on its prognostic value of esophageal cancer.

Under this circumstance, the definition of lymph node ratio (LNR) emerges to describe the lymph node status of patients more accurately. Until now, the association between low LNR and improvement in survival for patients has been proved in various malignancies, such as gastric cancer [19, 20] and breast cancer [14], as we have mentioned before. Meanwhile, LNR was also proposed to be an independent prognostic factor in gallbladder cancer [21] and pancreatic cancer [22]. A longitudinal study from a single cancer center in Taiwan found that high LNR was an independent poor prognostic factor in colorectal cancer [23].

In the previous studies, concerns were focused on comparing the accuracy or advantage between N stage and LNR in predicting prognosis of patients with esophageal cancer. Some studies [24] examined whether involvement of a higher number of lymph nodes was associated with worse survival among esophageal cancer patients. They found that disease-specific survival rates decreased with higher LNR and a higher LNR was independently associated with worse disease-specific survival. A prospective study of 224 patients was conducted to assess predictors of survival after esophagectomy for esophageal and gastroesophageal junction malignancy [25]. In the present study, our results showed a significant decrease in survival as the percentage of positive nodes increased. LNR was pointed to predict survival more accurately than current staging systems. The prognostic value of LNR in 700 esophageal squamous cell carcinoma (ESCC) patients after tri-incisional esophagectomy without neoadjuvant therapy was evaluated by making comparisons with N categories in the UICC/AJCC classification system (seventh edition), and the researchers concluded that LNR was an independent prognostic factor after tri-incisional esophagectomy, regardless of the number of retrieved lymph nodes [26]. They demonstrated that LNR might reduce stage migration and had more potential for predicting patient outcomes in ESCC.

In the present study, we evaluated the prognostic value of LNR in esophageal cancer. We divided the whole cases into two groups based on their LNR for survival analysis. We found that patients with LNR higher than 0.2 had a better OS and DFS than patients with LNR lower than 0.2. Our result was consistent with many previous studies, which clarified the significance of LNR in overall survival and disease-free survival of patients in esophageal cancer. One of the selected criteria of our study was esophageal cancer patients not receiving any therapy before esophagectomy. We intended to investigate the long-term outcome of patients after esophagectomy, and preoperative therapy can affect the outcome after esophagectomy to a certain extent. Preoperative therapy helps in downstaging esophageal cancer, thus enhancing the chance of curative resection and improving the prognosis of patients with esophageal cancer [27, 28]. So we excluded patients who had preoperative therapy from our study to eliminate this effect.

However, Mariette C and his colleagues [29] retrospectively analyzed 536 esophageal cancer patients taking into account neoadjuvant therapy and lymphadenectomy extent. Two hundred seventy-six patients who received neoadjuvant chemoradiation were included in this study. They found that the prognostic role of both the number and the ratio of metastatic lymph nodes was maintained regardless of patients receiving neoadjuvant chemoradiation or not. Besides, they had additional conclusion that LNR was shown to be more accurate for inadequately staged patients (<15 examined lymph nodes), whereas the number of metastatic lymph nodes was pertinent for adequately staged patients (≥15 examined lymph nodes).

Coincidentally, we also hold the opinion that it is difficult to demonstrate whether LNR is better than the number of metastatic lymph nodes. We need to take the results in the exact context to evaluate. In the present study, we have figured out that both LNR and N stage were associated with DFS and OS in esophageal cancer. But after comparison of survival rates between patient subsets in either N classification, we found significant difference in OS and DFS rates between different LNR categories within the same N stages but not between different N stages within the same LNR category. These results confirmed that LNR showed better prognostic value than N stage for esophageal cancer. This can be explained by the fact that LNR, determined by both metastatic lymph nodes and retrieved lymph nodes, is recognized as a better tool for revealing the lymphatic status of EC patients, compared with N stage. Lymph node metastasis is closely related with the recurrence and survival of esophageal cancer. Therefore, closer relation was found between LNR and survival rates.

As we all know, lymphadenectomy should take various factors into consideration, such as completeness of the resection, procedure-related morbidity and mortality, and disease-free survival [30]. In addition, many factors could influence the number of retrieved lymph nodes [31]. Thus, N staging, which was determined only by the number of metastatic lymph node, could not reflect the status of lymph node metastasis as accurately as LNR.

In this study, LNR >0.2 was associated with poorer OS and DFS in esophageal cancer. However, our study did not provide the evidence that the lower the LNR gained, the better the prognosis achieved. It has been proved that postoperative complications occurred more frequency with increasing number of lymph nodes retrieved [32, 33]. In other words, optimal LNR can increase the likelihood of proper staging and improve patient outcome with minimal complications.

There are some limitations about our study. First of all, the cutoff point of LNR lacks an acknowledged value at present. We selected 0.2 as the cutoff value according to the statistic method from the previous studies. Moreover, the sample size was small since there were only 209 cases included. Large sample clinical analysis is required for further study.

In conclusion, LNR was recognized as an independent factor in both OS and DFS in esophageal cancer. Besides, LNR showed better prognostic value than N stage for esophageal cancer. Further studies are needed to verify whether LNR could be used to select the appropriate preventive measures for an individual with poor prognosis to improve outcomes.