Introduction

Lymph node metastasis is one of the most important prognostic factors for colorectal cancer patients with no distant metastasis who have undergone curative resection. According to the American Joint Cancer Commission/Union for International Cancer Control tumor-node-metastasis (TNM) classification, nodal disease is divided into two prognostic risk groups depending on the number of metastatic regional lymph nodes: N1 indicates 1-3 metastatic lymph nodes and N2 indicates 4 or more metastatic lymph nodes.¹ Various studies have shown that the number of metastatic lymph nodes increases with the number of lymph nodes removed.^2,3,4,5,6 The number of lymph nodes removed is associated with the width of the surgical dissection, the extent of pathological examination, and the characteristics of the tumor and patient.^{3,4,5,7,8,9,10,11,12,13} The minimum number of excised lymph nodes required to stage nodal disease is specified as 7-14 in the 6^th edition of the TNM classification and as 10-14 in the 7^th edition; however, the number of lymph nodes recommended for removal is not taken into account when determining node-negative disease or in grouping based on the number of metastatic lymph nodes, and classification can be done even when fewer lymph nodes are removed.^1,14,15 Several studies accept the minimum lymph nodes harvest as 12.^{8,9,12,16,17,18,19,20} Lymph node ratio (LNR), (the ratio of metastatic lymph nodes to total lymph nodes removed) is a parameter that does not take into account the minimum number of lymph nodes that need to be removed, and numerous studies using a wide range of threshold values have reported its prognostic significance in colon and rectum cancers.

In the present study, we investigated the prognostic value of LNR in node-positive (stage III) colorectal cancer patients and whether it could contribute to the TNM classification.

Materials and Methods

Patients

Stage III colorectal cancer patients who were operated between January 1993 and December 2004 in our surgery unit, had no history of other malignancies, and underwent curative resection were evaluated. Nineteen patients who received neoadjuvant therapy, 10 patients with synchronous colorectal cancer, 11 patients with familial adenomatous polyposis, and 19 patients who died of complications in the early post-operative period were not included in the study. A total of 321 patients met the inclusion criteria for this study. Histological stage was grouped as low-grade (well or moderately differentiated) and high-grade (poorly differentiated, undifferentiated, mucinous, signet ring cell). All patients who received chemotherapy were given a 5-fluorouracil based regimen; chemotherapy was not administered to 35 patients whose general condition was poor or who rejected medical treatment. The clinicopathological data of all patients were collected prospectively.

Survival data of the patients were obtained from the oncology unit of our hospital or via phone calls to the patients or their relatives. The end point of the study was patient death. Cancer-specific survival (CSS) was defined as the time interval between surgery and mortality due to disease recurrence.^15,21 For 6 patients who developed a second malignancy, the date on which the second malignancy was diagnosed was accepted as the final follow-up date. For the 40 patients who died of causes unrelated to cancer, date of death was accepted as the final follow-up date.

This study was conducted in accordance with the principles of the Declaration of Helsinki and written informed consent was obtained from all patients prior to participation.

Statistical Analysis

In order to determine the LNR threshold value that would divide patients into two prognostic groups based on significantly different CSS rates, survival analyses were made with LNR values between 0.05 and 0.95 at increments of 0.05. Analyzes were performed on the entire series and subgroups including patients with <12 lymph nodes removed, ≥12 lymph nodes removed, TNM N1 disease, and TNM N2 disease.

CSS curves of the patient groups were calculated and rendered using the Kaplan-Meier method, and compared with the log-rank test. The relative significance of prognostic features was investigated using the Cox proportional hazards model. All comparisons were two-tailed. P values less than 0.05 were considered statistically significant. All statistical analyses were performed using SPSS version 17.0 (SPSS, Inc., Chicago, Illinois, USA).

Results

The clinicopathological characteristics and treatment specifications of the patients are shown in Table 1. One hundred fifty-one patients (47.1%) had fewer than 12 lymph nodes removed and 170 patients (52.9%) had 12 or more lymph nodes removed. TNM node class was stage N1 in 191 patients and N2 in 130 patients. One hundred ninety patients died of colorectal cancer by the end of the study in June 2013. The average follow-up time for surviving patients was 120.1 months.

Among the entire patient population, the most significant LNR threshold value that divided patients into low-risk (213 patients) and high-risk (108 patients) groups based on CSS was 0.40 (log-rank c²=66.216, p<0.001) (Figure 1), and LNR grouping had independent prognostic significance in multivariate Cox analyses (p<0.001) (Table 2). CSS was significantly different in patients with TNM N1 and N2 disease (log-rank c²=29.854, p<0.001) (Figure 2) and TNM node status had independent prognostic significance in the Cox analysis (p<0.001) (Table 3). When the LNR and TNM node grouping were combined in the Cox analysis, LNR maintained its independent prognostic significance (p<0.001), while TNM node grouping lost its prognostic significance (p=0.095) (Table 4).

In the patient group with <12 nodes removed, the most significant LNR threshold value discriminating low-risk (96 patients) and high-risk (55 patients) groups based on CSS was 0.40 (log-rank c²=42.911, p<0.001) (Figure 3). Including both LNR and TNM node class in the Cox analysis revealed that LNR maintained its independent prognostic significance (p<0.001), while TNM node class, which had prognostic significance in the univariate analysis, lost its prognostic significance (p=0.209). Among patients with ≥12 lymph nodes removed, the most significant LNR threshold value that divided patients into low-risk (117 patients) and high-risk (53 patients) groups was also 0.40 (log-rank c²=24.816, p<0.001) (Figure 4). When the LNR and TNM node class were both included in the Cox analysis, LNR grouping maintained its independent prognostic significance (relative risk=2.10, 95% confidence interval=1.23-3.57, p=0.006), while TNM node class, which had prognostic significance in the univariate analysis (log-rank c²=14.014, p=0.001) lost its prognostic significance (relative risk=1.48, 95% confidence interval=0.84-2.61, p=0.175).

In both the N1 and N2 patient groups, the most significant LNR threshold value separating low-risk (167 and 46 patients, respectively) and high-risk patients (24 and 84 patients, respectively) groups was 0.40 (log-rank c²=14.357 and log-rank c²=17.530 respectively, p<0.001).

In the group of patients with LNR ≤0.40, there was no significant difference (p=0.330) between the CSS rates of patients with N1 (167 patients) and N2 (46 patients) disease. There was also no significant difference in CSS between N1 (24 patients) and N2 (84 patients) patients with LNR >0.40 (p=0.132).

Discussion

The first study to investigate the prognostic significance of LNR in colon cancer was published in 2005. In that study, multivariate Cox analysis including both LNR and TNM node class in patients with ≥10 lymph nodes removed showed that TNM node class had no prognostic significance, while LNR was found to be the most significant prognostic factor.²² In a review of 16 studies, it was noted that LNR was superior to positive node number in the prognostic classification in stage III colon and rectal cancer patients who were not receiving neoadjuvant therapy.²³ LNR was also found to be a stronger prognostic factor than TNM node class in three population-based studies including large numbers of colon and colorectal cancer patients. In another population-based study involving colon cancer patients, LNR and TNM node class were each found to be independent prognostic factors.^24,25,26,27

In our series consisting of node-positive (stage III) colorectal cancer patients not receiving neoadjuvant therapy, an LNR of 0.40 had the highest discriminatory power between CSS prognostic groups in analyses of the entire series as well as subgroups of patients with <12 and ≥12 lymph nodes removed. In all three of these patient groups, TNM node class had prognostic significance in univariate analysis, but when the LNR and node class were both included in the Cox analysis, LNR continued to be a significant prognostic indicator while node class lost its prognostic significance. An LNR of 0.40 also divided patients into two prognostic groups with significantly different CSS rates in both the TNM N1 and N2 patient groups. On the other hand, there was no significant difference between the prognoses of N1 and N2 patients in both the group of patients with LNR ≤0.40 and those with LNR >0.40. These findings indicate that the LNR is a stronger prognostic parameter than TNM node class in colorectal cancer.

To date, studies have reported varying LNR categories to divide node-positive colon and rectal cancer patients into prognostic groups. LNR categories have been used to separate patients into five, three, or two prognostic groups.^{6,12,16,18,19,20,21,22,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42} All of these studies determined that LNR has independent prognostic significance. Studies which separated patients into 2 prognostic groups, as in our study, used LNR threshold values of 0.16, 0.17, 0.18, 0.20, 0.25, 0.30, and 0.60.^{6,20,37,38,39,40,41,42}

Some studies have evaluated the prognostic significance of LNR within patient subgroups created based on number of lymph nodes removed. LNR was shown to have prognostic significance in groups with <10 and ≥10 nodes removed and with <12 and ≥12 nodes removed.^{12,16,18,19,20,24} Different authors have reported that LNR had no prognostic significance below but did show prognostic value above lymph node harvest thresholds of 10, 12, and 13.^22,36,42

In some studies, the LNR has distinguished patients with TNM N1 and N2 disease into groups with significantly different prognoses.^{19,24,33,37,38} Another study showed that the LNR separated N2 patients into prognostic groups with significantly different survival rates, but could not do the same for N1 patients.²² Some studies reported no difference in survival rates between patients with N1 and N2 disease in patient subgroups created according to LNR.^19,33

As in our study, numerous authors have reported that in multivariate Cox analysis including LNR and TNM node class, only LNR maintained its independent prognostic significance.^{6,12,19,30,33,34,35,36,37,38,39,40,41,42} In one study, node class also maintained its independent prognostic significance in the Cox analysis, but LNR was found to be a stronger prognostic factor.²⁹

LNR is superior to TNM node class in separating node-positive colorectal cancer patients into prognostic groups, and may be useful in planning adjuvant therapy. However, there is a need for prospective clinical studies including sufficient numbers of patients to determine whether LNR can be used in place of or in combination with TNM node class, as well as to determine a standard LNR threshold value instead of the wide range of LNR threshold values that have been used previously.

Ethics

Ethics Committee Approval: Ethics Committee approval was not obtained because there was no ethics committee at that time. This study was conducted in accordance with the principles of the “Declaration of Helsinki”.

Informed Consent: Consent form was filled out by all participants.

Peer-review: Internally peer-reviewed.

Authorship Contributions

Surgical and Medical Practices: N.D., Concept: N.D., Design: N.D., Data Collection or Processing: Z.C.Ç., S.H., Analysis or Interpretation: N.D., Literature Search: Z.C.Ç., S.H., Writing: N.D.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.