A retrospective review was conducted between 2002 and 2015 at Manisa Celal Bayar University, and patients with a diagnosis of laryngeal benign, premalignant and malignant neoplasms were enrolled. The exclusion criteria were as follows: (i) patients with second primary cancer, (ii) patients with chronic inflammatory/autoimmune disorder, (iii) patients with malignant neoplasms other than squamous cell carcinoma. Hence, the study population included three groups: (i) patients with benign laryngeal neoplasms, (ii) patients with premalignant neoplasms, and (iii) patients with malignant neoplasms or LC. All cases were evaluated by clinical and radiological examinations such as ultrasonography, computed tomography or magnetic resonance imaging. TNM stages of patients with LC were determined according to the criteria of the 7th edition of “American Joint Committee on Cancer” Cancer Staging Manual established in 2009.

All samples for complete blood count were collected 5–10 days before the surgical biopsy/excision, which was performed under general anesthesia. Therefore, none of the patients had any sign of infection. The NLR value was measured by dividing the complete neutrophil count (CNC) to complete lymphocyte count (CLC), and the value of dNLR was determined by the formula of CNC/Complete white blood cell count (CWBCC)-CNC.

All statistical analysis was performed using SPSS v. 20.0 for Windows (Chicago, IL, USA). The mean values and standard deviations were calculated according to sociodemographic and clinicopathological characteristics of study population. Kolmogorov Smirnov and Shapiro–Wilk's tests demonstrated an abnormal distribution of NLR and dNLR values. Therefore, Mann–Whitney U test was performed for the comparison of two groups; and Kruskal–Wallis test was used for the comparison of three or more groups. In addition, the study population was dichotomized according to cut-off values of NLR and dNLR (low vs. high). The cut-off values were 4 for NLR and 2 for dNLR depending on clinical studies with high evidence level.5,6 Thereby, the diagnostic and predictive role of NLR and dNLR was also analyzed using Chi-Square test.

In this study, three endpoints (5 year: OS, disease free survival [DFS] and locoregional recurrence free survival [LRFS]) were selected for the assessment of oncological outcomes. All time to events were measured from the date of initial treatment modality. Overall survival was defined as the time from diagnosis to last follow-up or death for any cause. Disease-free survival was defined as the time from diagnosis to any sign of disease, locoregional recurrence (LR) and/or DM, last follow-up or death for any cause. Locoregional recurrence free survival was defined as the time from treatment to any sign of LR or last followup. Survival analysis was performed by Kaplan–Meier method and variables were compared using log-rank test. A Cox proportional hazard regression model was created in order to determine the effect of variables on oncological outcomes for both uni- and multivariate analysis. The variables with a probability value of <0.1 in the univariate analysis were selected for multivariate proportional Cox hazards regression analysis. Thereafter, only the variables with significance retained in the model using stepwise backward elimination; thereby independent prognostic factors were determined. The results were expressed as hazard ratios (HRs) with 95% confidence interval (CI). A value of p<0.05 was considered statistically significant.

The study population included 229 patients with laryngeal neoplasms. The descriptive statistics about patient, clinical, histopathological and treatment characteristics were presented in Table 1. The number of patients with benign, premalignant and malignant neoplasms was 30 (17 vocal fold polyps and 13 vocal fold nodules) (13.1%), 13 (9 mild dysplasia and 4 moderate dysplasia) (5.7%), 186 (81.2%), respectively. The majority of cases were male (221 cases, 96.5%). The median age of patients at the time of treatment was 59.0 (range 31–88) years (<65 years, 152 cases, 66.4%). A vast majority of laryngeal malignant neoplasms were glottic (82 cases; 44.1%) and supraglottic (79 cases, 42.5%). Transglottic larynx cancer was found in 24 cases (13.4%). Well, moderately, poorly differentiated and unidentified LC were seen in 25 (13.4%), 95 (51.1%), 15 (8.1%) and 51 (27.4) patients, respectively. Ninety-eight patients (52.7%) were advanced-stage LC, Stage III: 40 cases (21.5%); Stage IV: 58 cases (31.2%) and 88 patients (47.3%) were early-stage LC, Stage I: 48 cases (25.8%); Stage II: 40 cases (21.5%). thyroid cartilage invasion (TCI) was detected in 40 cases (21.5%). The distribution of T-stage I, II, III and IV were 26.3% (49 cases), 22% (41 cases), 28% (52 cases), 23.7% (44 cases), respectively. Lymph node metastasis (LNM) was found in 42 cases (22.6%), N1: 12 cases (28.5%); N2: 30 cases (71.5%) and extracapsular nodal spread was detected in 26 patients (61.9%). Seven patients (3.8%) with LC had DM at the time of diagnosis. Patients with early-stage LC were mainly treated with radiotherapy (57 cases, 64.8%). On the other hand, most of the patients with advanced-stage LC had surgical treatment and adjuvant radio±chemotherapy (67 cases, 68.4%). Ninety-eight patients (52.7%) were treated with surgery±adjuvant therapy and 88 cases (47.3%) with non-surgical modalities. The mean value of follow-up was 39.5 (1–107) months. During follow-up, LR and DM were detected in 33 cases (18.4%) and 19 cases (10.6%), respectively. Sixty-five patients (34.9%) died during the follow-up period.

The distribution, mean values and standard deviations of NLR and dNLR are presented in Fig. 1. The mean values of NLR and dNLR were relatively high in patients with LC; however, the difference was not statistically significant when patients with laryngeal benign, premalignant and malignant neoplasms were compared (p=0.068 and p=0.087) (Fig. 1). Similarly, no statistically significant difference was determined between patients with LC and other groups when the study groups compared according to low versus high NLR and dNLR values (p=0.232 and p=0.399) (Table 1). On the other hand, high NLR and dNLR values were highly correlated with supraglottic/transglottic, advanced stage LC and presence of TCI, LNM and DM (Table 1) (Fig. 1).

Figure 1.

The distribution, mean values, standard deviations and p-values of NLR and dNLR according to sociodemographic and clinicopathologic characteristics.

(0.61MB).

Five-year OS, DFS and LRFS of patients with LC were 61.3%, 62.3% and 61.8%, respectively. Patients with NLR≥4 had significantly poor oncological outcomes when compared with NLR<4; 5 year (OS: 69.0% vs. 31.1%, p<0.001; DFS: 70.0% vs. 32.7%, p<0.001; LRFS: 69.7% vs. 32.0%, p<0.001) (Fig. 2A–C). However, no statistically significant difference between low and high dNLR values was determined for 5 year: OS (65.4% vs. 54.8%, p=0.160); DFS (66.6% vs. 55.5%, p=0.234) and LRFS (66.3% vs. 54.8%, p=0.231) (Fig. 2D–F). Moreover, multivariate analysis using stepwise backward elimination demonstrated that NLR was an independent prognostic factor for 5 year: OS (HR=2.396, 95% CI 1.408–4.077; p=0.001); DFS (HR=2.246, 95% CI 1.322–3.816; p=0.006) and LRFS (HR=2.210, 95% CI 1.301–3.753; p=0.003) (Table 2).

Figure 2.

Log rank analysis of NLR and dNLR for 5 year OS, DFS and LRFS.

(0.17MB).

This study demonstrated that both NLR and dNLR values were not statistically different between patients with benign, premalignant and malignant neoplasms, even though high values were detected in patients with LC (Fig. 1). In contrast, Kum et al. and Duzlu et al.27,28 reported a statistically significant difference between patients with LC and other groups when the NLR values were compared. The study populations might be the major reason for this difference. It is obvious that patients with advanced-stage LC had remarkably high NLR and dNLR values when compared with patients with early-stage LC (Fig. 1). Inconsistent with our data, Wong et al.30 reported significantly higher values of NLR in patients with advanced-stage LC. As a matter of fact, a statistically significant difference might be determined when a study population that predominantly enrolled patients with advanced-stage LC.

In this study, standardized cutoff values were also preferred in order to stratify the study population and increase the quality of assessment. A cutoff value of 4 for NLR and 2 for dNLR were particularly selected depending on recent literature.5,6 A systematic review and meta-analysis, in which a cutoff value of 4 was selected, reported that patients with NLR≥4 had poor OS in several solid tumors.5 Moreover, Proctor et al.6 evaluated the prognostic role of NLR and dNLR in several cancers including head and neck cancers; and emphasized that optimal thresholds for NLR and dNLR were 4 and 2, respectively. Of note, we were unable to demonstrate a statistically significant difference between patients with benign, premalignant and malignant neoplasms even after the stratification of the study population according to abovementioned cut-off values (Table 1).

In patients with LC, the assessment of TCI, LNM, and DM is of utmost important for accurate clinical staging and treatment selection. Our study demonstrated that both NLR and dNLR were highly correlated with TCI, LNM, T stage, N stage, and M stage (Table 1) (Fig. 1). Similarly, Liu et al.35 determined a statistically significant relationship between NLR value and tumor size in patients with thyroid cancer. In our study, high NLR and dNLR values were also detected in patients with transglottic and supraglottic LC when compared with glottic LC. As a matter of fact, patients with transglottic/supraglottic tumors generally have locoregionally advanced LC and poor survival when compared with patients with glottic tumors.36 Therefore, all of the aforementioned results demonstrated that both NLR and dNLR are likely to be associated with local tumor extension, regional and distant disease, and may indicate the severity of larynx cancer.

In this study, patients with high NLR values (≥4) had a poor prognosis (Fig. 2A–C). Multivariate analysis also determined that NLR was an independent prognostic factor, and more than two-fold increase in the risk of mortality was determined in patients with high NLR value. In the literature, only a few numbers of clinical studies evaluated the prognostic role of NLR in patients with LC, and all of them used different cut-off values for risk stratification (Table 3). In this study, cut-off values of 4 for NLR and 2 for dNLR were particularly preferred due to the abovementioned scientific reasons. Inconsistent with our data, Tu et al.29 determined that NLR was a prognostic factor for both OS and DFS. In addition, Zeng et al. detected a statistically significant relationship between high NLR values and both OS and progression free-survival in patients with locoregionally advanced LSCC treated with chemoradiotherapy.31 Interestingly, Wong et al.30 reported that high NLR value was only correlated with poor OS, not with DFS. Recently, Kara et al. reported that NLR predicted local recurrence and decreased PFS.31 In addition, Wang et al.33 and Hsueh et al.34 showed the prognostic role of NLR in LC (Table 3).

In our study, no statistically significant difference was determined in oncological outcomes when a comparison according to high and low dNLR values was performed (Fig. 1). In addition, uni- and multivariate analysis demonstrated that a high dNLR value did not lead to an increase in the risk of mortality (Table 2). Similarly, Dirican et al.37 determined that NLR, not dNLR, was a prognostic factor for patients with breast cancer. In addition, Proctor et al.6 emphasized that NLR was a better biomarker to demonstrate prognosis in patients with cancer when compared with dNLR.