Lymph Node Ratio as a Prognostic Factor in Rectal Cancer Cases with Preoperative Chemoradiotherapy and Curative Resection

The number of metastatic lymph nodes (LNs) has been demonstrated to be one of the most important prognostic factors in colorectal cancer (Greene et al., 2002). However, the number of metastatic LNs is not only related to the severity of disease, but also depends on the number of retrieved LNs. For this reason, lymph node ratio (LNR; number of metastatic LNs/number of harvested LNs) has been studied in colorectal cancer to complement the current staging system for more precise prediction of patient prognosis (Peschaud et al., 2008; Rosenberg et al., 2008; Kim et al., 2009; Moug et al., 2009). These studies have found that, LNR is not only an important prognostic indicator, but also a more accurate stratification system than the current metastatic lymph node number-based staging system in colorectal cancer (Peschaud et al., 2008; Rosenberg et al., 2008; Kim et al., 2009; Moug et al., 2009).


Introduction
The number of metastatic lymph nodes (LNs) has been demonstrated to be one of the most important prognostic factors in colorectal cancer (Greene et al., 2002). However, the number of metastatic LNs is not only related to the severity of disease, but also depends on the number of retrieved LNs. For this reason, lymph node ratio (LNR; number of metastatic LNs/number of harvested LNs) has been studied in colorectal cancer to complement the current staging system for more precise prediction of patient prognosis (Peschaud et al., 2008;Rosenberg et al., 2008;Kim et al., 2009;Moug et al., 2009). These studies have found that, LNR is not only an important prognostic indicator, but also a more accurate stratification system than the current metastatic lymph node number-based staging system in colorectal cancer (Peschaud et al., 2008;Rosenberg et al., 2008;Kim et al., 2009;Moug et al., 2009). Preoperative chemoradiotherapy (CRT) followed by total mesorectal excision (TME) is the current standard of care for patients with T3 or T4 tumors and/or positive lymph nodes (Kapiteijn et al., 2001;Sauer et al., 2004). However, some studies have demonstrated that the number of harvested LNs is significantly decreased in rectal cancer patients received preoperative CRT (Rullier et al., 2008;Doll et al., 2009;Wang et al., 2009). The number of harvested LNs in rectal cancer patients treated with preoperative CRT is frequently fewer than 12, as recommend by the American Joint Committee on Cancer (AJCC) and the National Cancer Institute (Goldstein et al., 1996). Therefore, LNR, which takes both the number of positive LNs and harvested LNs into consideration, may serve as a better prognostic indicator in rectal cancer patients received preoperative CRT. The aim of this study is to evaluate the impact of LNR on prognosis in ypNpositive rectal cancer patients.

Materials and Methods
We performed a retrospective review of a prospectively entered database, from January 1999 to December 2011, a total of 140 consecutive patients with primary rectal cancer treated with preoperative CRT followed by curative resection were pathologically diagnosed with nodepositive (ypN-positive) rectal cancer at our institution. Nine patients were excluded from this study (five were lost to follow-up, one was died of postoperative complication, three refused to receive postoperative chemotherapy). The remaining 131 patients were included in the analysis.
All the patients underwent digital rectal examination, colonoscopy, abdominal and pelvic computed tomography (CT), and chest X-ray for clinical staging. Transrectal ultrasonography was performed in 102 (77.9%) patients, and 118 (90.1%) patients received pelvic magnetic resonance imaging for preoperative staging. Preoperative CRT was delivered to patients who had a clinical stage of T3 or T4 and/or positive lymph nodes.
The details on radiotherapy have been previously published (Jin et al., 2006). In brief, a total dose of 50 Gy was delivered in 2.0-Gy daily fractions to the pelvic area. Preoperative chemotherapy was initiated on the first day of radiotherapy. Capecitabine was administered concurrently with radiotherapy at a dose of 1, 600 mg/ m2/day for 35 days After preoperative CRT, all patients underwent curative resection, and TME principle was followed for every patient. The median interval between preoperative CRT and surgery was 7 weeks (range 6-8 weeks).
Postoperative specimens were examined by at least two pathologists specialized in colorectal cancer. When less than 12 lymph nodes were found, re-examination was performed by a third pathologist. The pathologic stage of the tumor was determined according to the American Joint Committee on Cancer (seventh Edition) staging system. About 3 to 6 weeks after completing the surgery, all the patients received postoperative chemotherapy. Two different chemotherapy regimens were used: (1) capecitabine, and (2) oxaliplatin, leucovorin, and 5-fluorouracil.
After hospital discharge, patients were suggested to visit the doctors every 3 month within first 2 years and every 6 month thereafter. During each follow-up, patients received a series of evaluations, including digital rectal examination, complete blood count, liver function test, and carcinoembryonic antigen (CEA) level test. Abdominal and pelvic computed tomography (CT), and chest X-ray were conducted every 6 months after surgery. Colonoscopy was performed per year after surgery.

Statistical analysis
Continuous variables are expressed as median and range, and were analyzed with the Student's t-test, while categorical ones are expressed as numbers with percentages, and were analyzed by chi-square test or Fisher's exact test when appropriate. Overall survival (OS) was defined from the date of operation to the date of death. Disease-free survival (DFS) was defined as the time from operation to local recurrence, metastasis, or death. Kaplan-Meier method was used to analyze survival of patients, and comparisons were analyzed by log-rank test. Cox's proportional hazards model was used for multivariate analysis, adjusted hazard ratios (HRs) and their 95%CIs were calculated. All statistical tests were two-sided, and a P value of less than 0.05 was considered statistically significant. Data were analyzed by Statistical Package for the Social Science (SPSS) 18.0 for Windows (SPSS Inc. Chicago, IL, USA).

Subgroup analysis according to number of harvested LNs and ypN classification
As we found LNR was an independent prognostic factor for OS and DFS, we next evaluated the prognostic value of LNR in different subgroups (Table 4). Firstly, the prognostic impact of LNR was analyzed according to the number of harvested LNs. Increased LNR was significantly associated with worse OS and DFS in patients with <12 harvested LNs, and as well as in those≥12 harvested LNs (p<0.05). In order to find out whether LNR had a prognostic impact in each ypN classification, we performed a subgroup analysis based on LNR in patients with ypN1 and ypN2 stages, respectively. In the ypN1 subgroup, increased LNR was significantly associated with reduced OS and DFS rates (p<0.001). In the ypN2 subgroup, increased LNR also associated with worse OS and DFS, but the difference did not reach statistically significant (p>0.05). According to LNR, patients in ypN1

Prognostic factor for survival
In univariate analysis (Table 2), LNR, ypT stage, and ypN stage were significantly associated with OS and DFS. Other clinicopathologic characteristics, such as age, gender, and type of surgery, were not significantly associated with OS or DFS. In multivariate analysis (Table  3), LNR was identified as an independent prognostic factor for both OS (HR=3.778;p=0.001) group could be divided into two groups (N1R1 vs N1R2). Patients in N1R2 group had similar 5-year OS (53.9% vs 52.8%; p=0.930) and DFS (28.8% vs 34.1%; p=0.738) rates to ypN2 group (Figure 2).

Discussion
Preoperative CRT followed by total mesorectal excision is now standard treatment for rectal cancer patients with T3 or T4 tumors and/or positive lymph nodes (Kapiteijn et al., 2001;Sauer et al., 2004;Lee et al., 2013). Adequate examination of the regional lymph node plays a vital role in the prediction of patient prognosis, as lymph node involvement after preoperative CRT is the most important prognostic factor in these patients (Kim et al., 2006;Chang et al., 2009). According to current TNM staging system, at least 12 LNs are needed for accurate nodal staging. Many factors influence the number of lymph nodes retrieved, including tumor (size, stage) (Baxter et al., 2005), and the patient (age, sex) (Thorn et al., 2004;Gao et al., 2013), and neoadjuvant CRT (Rullier et al., 2008;Doll et al., 2009;Wang et al., 2009). Recent studies demonstrated that the total number of retrieved LNs was decreased due to preoperative chemoradiation, probably because of lymph node atrophy, fibrosis and lymphocyte depletion caused by radiotherapy or/and chemotherapy, and the number of harvested LNs was frequently less than12, despite the maintenance of vigorous surgical standards (Rullier et al., 2008;Doll et al., 2009;Wang et al., 2009;Lee et al., 2012). Lee et al. (2012) found that <12 LNs were harvested in 30.5% patients after preoperative CRT. Similarly, we found that <12 LNs were retrieved in 32.8% patients.
In the 7th AJCC staging system, nodal staging system for colorectal cancer is solely based on the number of metastatic lymph nodes. In order to complement the current staging system for more precise prediction of patient prognosis, LNR, incorporating both the number of harvested LNs and positive LNs in one prognostic value, has shown to be an important prognostic indicator for colorectal cancer patients received postoperative adjuvant treatment (Peschaud et al., 2008;Rosenberg et al., 2008;Kim et al., 2009;Moug et al., 2009). However, studies of prognostic value of LNR in rectal cancer patients treated with preoperative CRT are limited.
In the current study, we investigated the prognostic value of LNR in 131 rectal cancer patients treated with preoperative CRT followed by curative resection. In multivariate analysis, LNR showed to be an independent prognostic factor for both OS and DFS. In subgroup analysis, LNR was also significantly associated with OS and DFS in patients with <12 harvested LNs, and as well as in those≥12 harvested LNs. Moreover, LNR had a prognostic impact on both OS and DFS in patients with N1 stage. Subgroups of the ypN1 stage divided by the cut-off value 0.2 of LNR showed a different prognosis. ypN1 patients with an LNR>0.2 had a significantly worse 5-year OS (53.9% vs 92.9%; p<0.001) and DFS (28.8% vs 91.1%; p<0.001) than ypN1 patients with an LNR≤0.2. Although previous studies indicate that lymph node involvement after preoperative CRT is the most important prognostic factor (Kim et al., 2006;Chang et al., 2009), our results demonstrated the limitations of the currently used ypN staging system, which, in fact, was not originally designed to stratify patients after neoadjuvant CRT. Our results showed that LNR could stratify patient prognosis more accurately than pN staging. The superior stratification power of LNR over ypN stage might be due to the fact that patient grouping by LNR is less influenced by the total number of harvested LNs than patient grouping by ypN stage.
There are several limitations in our study. Because this study was not prospectively designed, it is subject to potential bias. The cutoff value of LNR is different from other studies (Kang et al., 2011;Klos et al., 2011;La et al., 2013;Madbouly et al., 2013;Nadoshan et al., 2013). In fact, the cutoff value varies among different studies, ranged from 0.07 to 0.6 (Madbouly et al., 2013). However, all of these studies find that LNR is a powerful predictor of outcome and sometimes even more powerful than nodal status. In accordance with our study, Nadoshan et al. (2013) demonstrated that LNR >0.2 was a significant prognostic factor for survival in patients with stage III rectal cancer undergoing pre-operative chemoradiotherapy. Nevertheless, Madbouly et al. (2013) showed that a cutoff value of 0.375 was determined to be the most accurate predictive point. These different cutoff values could be explained by the disproportion of patient grouping according to ypN stage, LNR, and mean number of LNs retrieved. The inconsistent cutoff value of LNR remains one of the limitations of validation in clinical practice. Prospective studies are needed to define the LNR cutoff value that will optimize patient prognosis in stage III rectal cancer.
In conclusion, LNR is an independent prognostic factor for OS and DFS in ypN-positive rectal cancer patients, both in patients with <12 harvested LNs, and as well as in those≥12 harvested LNs. In addition, LNR has a prognostic impact on both OS and DFS in patients with N1 staging. LNR provides better prognostic value than pN staging. Therefore, LNR should be used as an additional prognostic indicator in ypN-positive rectal cancer patients.