Significance of Neutrophil-Lymphocyte Ratio (NLR) as a Prognostic Factor in Stage II Colorectal Cancer

A total of 124 patients with stage II CRC who underwent tumor resection with anastomosis between April 2006 and March 2013 were enrolled in this analysis. The median follow-up for all patients was 2179 (range, 178–3650) days. The preoperative characteristics of the patients included in this study are summarized in Table 1.

This study was approved by the Ethics Committee of Tokyo Women's Medical University (approval number 4750-R). We followed the retrospective observational research information disclosure procedure for obtaining informed consent from research subjects.

NLR was calculated by dividing the number of neutrophils by the number of lymphocytes obtained from complete blood counts (Fig. 1).⁶  Patients were stratified into 2 groups according to NLR cutoff values (high NLR, ≥ cutoff value; normal NLR, ≤ cutoff value), which were obtained from receiver operating characteristic (ROC) curves.

Disease-free survival (DFS) and overall survival (OS) were compared between the 2 groups. Univariate and multivariate analyses were performed using the Cox proportional-hazards model to determine whether NLR was a prognostic factor associated with DFS and OS.

The following patient-related factors and clinicopathologic characteristics were categorized, except NLR: sex (male versus female), age (≥80 versus <80 years), prognostic nutritional index⁶  (PNI; ≤40 versus >40), controlling nutritional status⁷  (CONUT; 0–1 versus ≥2), modified Glasgow prognostic score⁸  (mGPS; 0–1 versus 2), body mass index (BMI; <22.0 versus ≥22.0 kg/m²), performance status (PS; 0 versus ≥1), American Society of Anesthesiologists (ASA) classification (1–2 versus 3), serum carcinoembryonic antigen (CEA; ≤5.0 versus >5.0 ng/mL), depth of tumor invasion⁹  (T3 versus T4), tumor histology⁹  (papillary adenocarcinoma or tubular adenocarcinoma versus other), and tumor location (rectum versus colon).

All statistical analyses were conducted using JMP Pro 11.0 (SAS Institute Inc., Cary, North Carolina). The ROC curve and the area under the ROC curve (AUC) were used to determine the optimal preoperative NLR cutoff values for predicting recurrence.

For the analysis of DFS and OS, we compared the 2 patient groups, classified based on preoperative NLR. Survival curves were estimated using the Kaplan–Meier method and compared using the log-rank test. Univariate and multivariate analyses and the Cox proportional-hazards model were used to compare survival times. Hazard ratios (HRs) and 95% confidence intervals (CIs) were reported. P < 0.05 was considered statistically significant.

The median NLR for all subjects was 2.6 (range, 0.4–12.7). The AUC for NLR was 0.81 (95% CI: 0.70–0.91). The sensitivity and specificity were 77.8% and 84.0%, respectively. The optimal NLR cutoff value for predicting recurrence obtained from the ROC curve analysis was 4.0. Thirty-one patients had a high NLR, and 93 patients had a low NLR (Fig. 2).

Of the 124 patients included in this study, 18 (14.5%) experienced recurrence during follow-up, with a median of 559 (range, 181–2952) days. Among them, 2 patients developed recurrence after 5 years. The 5-year DFS rate of patients with a high NLR was significantly lower than that of patients with a low NLR (62.7% versus 95.6%, respectively; P < 0.0001; Fig. 3).

Twenty-four (19.4%) of the 124 patients included in this study died during follow-up. The median duration was 1550 (range, 199–3200) days. Thirteen patients died after 5 years; 8 patients (33.3%) died from cancer. The cumulative 5-year OS rate of patients based on the NLR cutoff value of 4.0 was a significant determinant of poor prognosis in the high NLR group (66.7% versus 95.5%, respectively; P < 0.0001; Fig. 4).

Univariate analysis showed that age > 80 years, NLR ≥ 4.0, PNI ≤4 0, CONUT score ≥ 2, mGPS of 2, PS ≥ 1, serum CEA level > 5.0 ng/mL, and depth of tumor invasion (T4) were associated with poor DFS in patients with stage II CRC.

In the multivariate analysis, NLR ≥4.0 (HR: 5.31, 95% CI: 1.30–21.7; P = 0.02), PS ≥1 (HR: 7.03, 95% CI: 1.81–27.3; P = 0.01), and depth of tumor invasion (T4) (HR: 4.81, 95% CI: 1.15–20.05; P = 0.048) were associated with poor DFS in patients with stage II CRC (Table 2).

Univariate analysis showed that age > 80 years, NLR ≥ 4.0, PNI ≤ 40, CONUT score ≥ 2, mGPS of 2, BMI < 22.0 kg/m², PS ≥ 1, and depth of tumor invasion (T4) were associated with poor OS in patients with stage II CRC.

In the multivariate analysis, age > 80 years (HR: 3.63, 95% CI: 1.35–9.77; P = 0.01), NLR ≥ 4.0 (HR: 4.39, 95% CI: 1.17–16.37; P = 0.03), and PS ≥ 1 (HR: 4.90, 95% CI: 1.43–16.83; P = 0.01) were associated with poor OS in patients with stage II CRC (Table 3).

Recent studies have shown that elevated systemic inflammation, which is a host response in patients with preoperative CRC, is associated with poor outcomes independent of tumor stage.^2,3  Among the parameters that reflect systemic inflammation in patients with preoperative CRC, NLR is considered a reliable marker because it demonstrates the balance between preneoplastic and antitumor immune status.^2–5  Furthermore, patients with a high NLR have lymphocytopenia and neutrophilic leukocytosis, which favor protumor inflammation and lead to poor outcomes.² 

Initially, we determined the NLR cutoff value to be 4.0 using ROC curve analysis. However, it must be noted that the NLR threshold value has not been clearly defined.^{2–5,10–15}  The cutoff values reported in previous studies ranged from 2.4 to 5.0 for stage II CRC, although an NLR cutoff value of 5.0 is frequently used for all stages.²  Despite the different NLR cutoff values mentioned above, several studies have shown that an NLR above the cutoff value in patients with CRC is associated with poor prognosis.^2,3,11–15  It has also been reported that the NLR cutoff value tends to be set higher in stages I to III and lower in stage IV CRC as NLR values are influenced by tumor-related factors.³  Hence, because of multiple similar observations,^3,11  a high NLR value of 4.0 in patients with stage II CRC was used as the cutoff value in this study.

We subsequently compared DFS and OS rates according to the NLR cutoff value of 4.0. Patients with a high NLR had a 5-year DFS rate of 62.7% and a 5-year OS rate of 66.7%, both of which were lower than those of patients with a low NLR (95.6% and 95.5%, respectively). This finding was consistent with previous studies of patients with stage IIA CRC.^10–13 

A high NLR indicates a higher neutrophil count compared with lymphocyte count in response to stress (e.g., infection and inflammation).^2,14,15  However, the association between a high NLR and poor prognosis is complicated, and the mechanism has not yet been fully elucidated.

A high NLR indicates lymphocytopenia, which may lead to a poor lymphocyte-mediated immune response to the tumor because of the reduction in the ratio of T4 (helper) to T8 (suppressor) cells.^14,15  A subsequent reduction in the number of lymphocytes impairs the hosts' antitumor immune response, thus conferring a poor prognosis.^14,15 

Conversely, an increase in neutrophil count promotes tumor vascularization and tumor growth and metastasis because of the release of proangiogenic chemokines and other factors.^2,14,15  Because the hosts' immune environment also plays an important role in CRC cancer metastasis, the loss of this balance, which leads to a high NLR, may contribute to poor prognosis.^2,14,15 

Finally, multivariate analysis revealed that a high NLR was an independent predictor of DFS in patients with stage II CRC, in addition to PS and depth of tumor invasion (T4). There is no direct evidence available on the relationship between PS and the recurrence rate of CRC. However, in the case of a poor PS, natural killer cell activity is low in advanced-stage cancers,¹⁶  which may reflect potential tumor progression. In the present study, in patients with stage II CRC, a poorer PS was associated with a lower OS rate.

The depth of tumor invasion (T4) was also an independent risk factor for DFS, a finding that is consistent with a number of previous studies.^15,17,18 

Ding et al¹² reported an association between NLR (cutoff value of 4.0) and recurrence-free survival in patients with stage IIA CRC. Furthermore, cutoff values of 2.6 in patients with stage II CRC¹¹  and of 2.0 in patients with stage II–III CRC¹⁹  have been reported as independent prognostic factors for DFS.

Multivariate analysis revealed that a high NLR is an independent predictor of OS in patients with stage II CRC, in addition to age > 80 years and PS. NLR has been reported in several studies to be an independent predictor of OS. Seung et al¹¹ reported a cutoff value of 2.6, even for patients with stage II CRC. Regarding the association between NLR and noncancer mortality rates in patients with stage II colon cancer, Hung et al¹⁵ reported that the noncancer mortality rate was lower in patients with an NLR < 5 (15.8%) and higher in patients with an NLR > 5 (23.9%). These results suggest that NLR is not only a predictor of cancer mortality but also a predictor of noncancer mortality.¹⁵ 

Conversely, the results on OS in this study demonstrated that 16 of 24 patients died during follow-up from causes other than cancer and that a high NLR was associated with poor OS. Because noncancer mortality was also high in this study, NLR may not be the only predictor of noncancer mortality. Nevertheless, NLR may be a predictor of cancer mortality because it is associated with OS.

Several reports suggest that older patients have poorer OS than DFS compared with younger patients.^15,19  Therefore, postoperative mortality in older adults is high because of the presence of comorbidities.²⁰  This is consistent with the results of this study, wherein older patients had worse OS than younger patients.

Furthermore, patients with a poor preoperative PS have several preoperative comorbidities, a high incidence of postoperative complications, and a long duration of hospitalization,²¹  which are important factors for predicting OS.^19–21  In the present study in patients with stage II CRC, a decrease in PS was associated with a poor OS.

This study has several limitations. First, this was a retrospective, single-center study; therefore, there is potential for bias in patient selection. Second, the cutoff value may require further assessment in future studies.

NLR is an inexpensive assessment method that can be measured at any facility. NLR may also be frequently measured in daily medical care for patients with CRC; hence, it has a utility value that is not present in existing markers. For these reasons, NLR may have great clinical applicability in the future.³ 

In conclusion, the present study identified preoperative NLR as an independent poor prognostic factor for DFS and OS in patients with stage II CRC undergoing curative resection. This inflammatory marker might be a useful biomarker in patients with stage II CRC.