Prognostic Role of Systemic Inflammatory Markers in Patients Undergoing Surgical Resection for Oral Squamous Cell Carcinoma

Background: A high platelet–lymphocyte ratio (PLR) is a marker of systemic inflammation and, together with the neutrophil–lymphocyte ratio (NLR), is associated with poor outcomes in several cancers. We investigated the prognostic value of PLR and other systemic inflammatory markers, such as NLR, systemic immune-inflammation index (SII), and systemic inflammation response index (SIRI), in oral squamous cell carcinoma (OSCC) patients undergoing surgical resection. Methods: We derived PLR, NLR, SII, and SIRI from a retrospective chart review of 269 consecutive OSCC patients. The complete blood count examined in the immediate preoperative period was used to compute PLR, NLR, SII, and SIRI. We analyzed the relationship between these systemic inflammatory markers and the clinicopathologic characteristics, disease-specific survival (DSS), and progression-free survival (PFS) of patients. Results: In the univariate analysis, high PLR and SII were significantly associated with worse DSS and PFS (all p < 0.05). In the multivariate analysis, PLR (HR 2.36, 95% CI 1.28–4.36 for DSS; HR 1.80, 95% CI 1.06–3.06 for PFS) was an independent predictor of survival outcomes. When PLR was analyzed as a continuous variable, the relationship between the outcome and preoperative PLR was not monotonically linear. In the subgroup analysis, PLR was more strongly associated with DSS and PFS in patients who were male, had stage III/IV OSCC, or had lymph node metastasis. Conclusion: Our data suggest that in OSCC patients, the pretreatment PLR is an independent predictor of DSS and PFS. The PLR is a readily available biomarker that will improve prognostication and risk stratification in OSCC.


Introduction
Oral squamous cell carcinoma (OSCC) is a carcinoma with squamous differentiation arising from the mucosal epithelium of the oral cavity and mobile tongue. The global incidence of oral cancer, the majority being squamous cell carcinoma, is approximately 3.5 million new cases per year, and it causes 1.7 million deaths per year. Oral cancer accounts for 2.0% of all cancers [1]. In Korea, oral cancer is the second most common cancer among head and neck cancers, and the incidence of oral cancer has been slightly increasing in recent decades [2,3]. The incidence rate is especially rising more steeply in the third-or fourthdecade age groups [4]. Smoking, drinking alcohol, lifestyle changes, the popularization of early diagnosis, and genetic factors could be the causes of such an increase [4,5]. Oral cancer is more common among men than women, and most common in the fifth and sixth decades [6]. The survival rate of oral cancer is approximately 50% [7]. Advancements in traditional treatment modalities, i.e., surgery, chemotherapy, and radiotherapy, have not

Statistics
The characteristics of the systemic inflammatory markers are shown as both medians and means. Student's t-test was used to compare continuous characteristics between the two groups. Pearson's test was used to analyze the correlation between two continuous variables. DSS was considered the period between surgery and the date of the last followup, or cancer-specific death. PFS was considered the period between surgery and the date of recurrence, locoregional progression, metastasis, or death. The DSS and PFS rates were analyzed using Kaplan-Meier survival curves and compared statistically with the log-rank test. Age, sex, and all variables with significant prognostic values in the univariate analysis were subjected to multivariate analyses using the Cox proportional hazards model. A twosided p < 0.05 was considered statistically significant. The potential nonlinear relationships between the continuous PLR and the survival outcomes were flexibly analyzed using a restricted cubic spline (RCS) with four knots [36]. The median value was used as a continuous predictor. Statistical analyses were conducted using SPSS 21.0 for Windows (IBM Corporation, Armonk, NY, USA) and R version 4.1.2.

Correlation between Inflammatory Markers and Clinical Factors
We examined the correlation between systemic inflammatory markers and clinicopathologic parameters. The mean WBC count was higher in patients with a depth of invasion >1 cm and advanced T and AJCC stages (1 and 2 vs. 3 and 4) (all p < 0.05) (Supplementary Table S1). Platelet levels were higher in the younger patient group (p = 0.0284), but other markers showed no difference between the two age groups (Supplementary  Table S1). There were no significant differences between the high and low NLR groups in the clinicopathologic parameters. High PLR was correlated with >1 cm depth of invasion and advanced T and AJCC stages (all p < 0.05). Likewise, advanced stage was correlated with high SII and SIRI (p = 0.0015 and 0.0131, respectively) ( Table 4).  The survival analysis results are shown in Table 5. The median follow-up period was 36 months (range 0~185 months). Of the 269 patients, 65 patients died during the follow-up period, and 93 patients experienced disease progression. The 2-year and 5-year DSS rates of the OSCC patients were 78.8% and 75.6%, respectively.  Among the systemic inflammatory markers, the survival analysis revealed poorer DSS and PFS for patients with high PLR and SII (all p < 0.05) (Figures 1 and 2). The 5-year DSS rates of the low vs. high PLR groups were 81.2% vs. 58.4% (p = 0.0004), and the 5-year PFS rates of the low vs. high PLR groups were 70.2% vs. 45.9% (p = 0.0002). NLR was associated with PFS (p = 0.0485), and SIRI showed no association with DSS or PFS (p = 0.012) (Figures 1 and 2).

Analysis of the Relationship between PLR and Survival According to Clinical Factors
We conducted subset analyses of the impact of PLR on survival according to selected clinicopathologic factors using forest plots. Notably, PLR was more strongly associated

Analysis of the Relationship between PLR and Survival According to Clinical Factors
We conducted subset analyses of the impact of PLR on survival according to selected clinicopathologic factors using forest plots. Notably, PLR was more strongly associated with DSS and PFS in patients who were male, had stage III/IV OSCC, or had lymph node metastasis (all p < 0.05) ( Figure 3). with DSS and PFS in patients who were male, had stage III/IV OSCC, or had lymph node metastasis (all p < 0.05) (Figure 3).

Nonlinear Association between PLR and Survival
Furthermore, PLR was studied as a continuous variable in univariate analysis. The RCS analysis showed a curvilinear, J-shaped association between the PLR and survival outcomes rather than a straight line (Figure 4). This result suggested a possible nonlinear association between the PLR and the risk of DSS (p-value for nonlinearity = 0.0424). We did not find statistical evidence for nonlinearity with the progression-free survival outcome. We estimated the mortality risk to reach a nadir PLR in the range of 100-120, with inverse associations below that range and positive associations above that range, although the magnitude of associations varied.

Nonlinear Association between PLR and Survival
Furthermore, PLR was studied as a continuous variable in univariate analysis. The RCS analysis showed a curvilinear, J-shaped association between the PLR and survival outcomes rather than a straight line (Figure 4). This result suggested a possible nonlinear association between the PLR and the risk of DSS (p-value for nonlinearity = 0.0424). We did not find statistical evidence for nonlinearity with the progression-free survival outcome. We estimated the mortality risk to reach a nadir PLR in the range of 100-120, with inverse associations below that range and positive associations above that range, although the magnitude of associations varied.
with DSS and PFS in patients who were male, had stage III/IV OSCC, or had lymph node metastasis (all p < 0.05) (Figure 3).

Nonlinear Association between PLR and Survival
Furthermore, PLR was studied as a continuous variable in univariate analysis. The RCS analysis showed a curvilinear, J-shaped association between the PLR and survival outcomes rather than a straight line (Figure 4). This result suggested a possible nonlinear association between the PLR and the risk of DSS (p-value for nonlinearity = 0.0424). We did not find statistical evidence for nonlinearity with the progression-free survival outcome. We estimated the mortality risk to reach a nadir PLR in the range of 100-120, with inverse associations below that range and positive associations above that range, although the magnitude of associations varied.

Discussion
The inflammatory response is greatly influenced by tumor manifestation [37,38]. Many cancers cause extensive disruption of hematopoiesis [11]. In particular, cell mediation is closely associated with tumor development, growth, and metastasis [39]. There are methods to measure the inflammatory response, including C-reactive protein, erythrocyte sedimentation rate, and peripheral blood cell count [40,41]. NLR, PLR, and monocyte-lymphocyte ratio (MLR) are measured from peripheral blood cell counts, and illustrate how much neutrophils, platelets, and monocytes are increased compared to lymphocytes. These changes represent the cell-mediated systemic inflammatory response [42]. In studies conducted over the last decade, the presence of an elevated NLR and PLR has been associated with poorer outcomes in different types of malignancies [14,42,43]. Some researchers termed systemic inflammatory responses "the tip of the cancer iceberg" [44], but this approach is underutilized in clinical practice.
Recently, new prognostic scores or indexes were created through a combination of nutrition index, performance index, or three or more peripheral blood cell counts [23,45]. SII and SIRI are recently suggested novel prognostic biomarkers, derived from a combination of the absolute neutrophil count, lymphocyte count, and monocyte count. The prognostic value of each has been suggested in patients with lung cancer [24]. In our study, high SII and SIRI were associated with poor DSS in patients with OSCC in the univariate analysis. They were promising prognostic biomarkers in OSCC, but they failed to be identified as independent prognostic factors in the multivariate analysis.
In OSCC, several studies have investigated the prognostic value of NLR and PLR [33,46]. In a meta-analysis of 10 studies, a high NLR was associated with a poor prognosis in patients with OSCC [33]. PLR was also an independent prognostic factor in previous studies [46][47][48], and PLR was superior in the studies by Tazeen et al. [46] and Rosculet et al. [48]. In concordance with previous studies, our data showed that PLR was more strongly associated with overall survival and PFS than NLR, SII, and SIRI in patients with OSCC. NLR, SII, and SIRI were associated with either DSS or PFS in the univariate analysis, but PLR remained the only significant prognostic indicator after adjustments in the multivariate analysis. The subset analysis revealed that PLR had a greater prognostic impact in patients with advanced disease (stage III and IV) than in those with localized disease (stage I and II) (HR = −0.01, 95% CI −1.32 to 1.32, p = 0.9932 vs. HR = 0.84, 95% CI 0.29 to 1.40, p = 0.0028). This may be explained by the fact that cancer cells interact little with the microenvironment and local inflammatory cells in the low stages, and subsequently elicit little systematic inflammation [11,39,49]. Our data suggest that more active surveillance or treatment may be required for patients with high PLR, especially for those with stage III or IV OSCC. We need to actively seek alternative therapies such as immunotherapy or molecular targeted therapy in this patient group because their prognosis is predicted to be poorer after surgery.
Furthermore, we analyzed PLR as a continuous variable as well as a dichotomized variable. Interestingly, the prognostic correlation of PLR did not show monophasic linearity. The J-shaped relationship between DSS and PLR indicates an unfavorable prognosis for not only the high PLR group but also the extremely low PLR group. A J-shaped association is often observed in epidemiologic relationships, such as the association of body mass index and mortality [50]. A few researchers demonstrated a similar nonlinear association between NLR and prognosis in patients with gastric, breast, and oral cancers in the United States general population [51][52][53][54][55]. The nonlinear relationship between the survival outcome and pretreatment PLR has never been demonstrated previously. Our data showed that PLR has a nonlinear prognostic pattern similar to NLR. The increased risk of DSS at extremely low PLRs might indicate the strong cancer-related disturbance of normal hematopoiesis in these patients and its effect on survival, or the pervasive effect of unknown underlying noncancer-related health conditions. Nevertheless, an explicit rationale for such a pattern has yet to be established. It is necessary to further research the nonlinear prognostic pattern of PLR and to consider how to employ the best PLR thresholds in clinical practice.
The underlying mechanisms of PLR as a prognostic marker have not been fully elucidated, but there are several explanations. Bodies of evidence have shown that platelet activation is a key biological process for cancer occurrence, progression, and metastasis [56,57]. The phenomena related to platelet activation through interaction with cancer cells are thrombocytosis and thromboembolism [58]. Thrombocytosis is often observed in patients with advanced malignancy, and thromboembolism occurs frequently in cancer patients, with lethal consequences [57]. Aggregated platelets could enhance tumorigenesis by releasing pro-angiogenic mediators within the tumor microvasculature [59]. Furthermore, platelets could influence the metastatic potential of cancer cells through several biological pathways, i.e., secretion of cellular growth factors, helping stable tumor cell adhesion to endothelial cells, and impeding cell-mediated immunity against tumor cells [59][60][61][62][63][64]. A newly discovered link between platelets and cancer cells is tumor-educated platelets [65]. Tumor-educated platelets were most recently found through liquid biopsy research. Tumoreducated platelets are functional cells with a distinct tumor-driven phenotype that are thought to acquire tumor-derived factors and undergo signal-dependent changes in RNA processing within blood circulation [65]. Tumor-educated platelets have been shown to participate in multiple steps of metastasis, leading to lethal consequences [66,67]. There is a hypothesis that microvesicles containing RNA and proteins taken up by platelets promote tumor growth and immune evasion [68,69]. The relationship between tumor-educated platelets and PLR elevation is unclear. Future research on this matter would be interesting and would help elucidate the biological mechanisms underlying PLR as a biomarker.
The strength of our study is that various, not a single, systemic inflammatory markers were analyzed and that this was performed in a homogenous, rich sample group. Exploration of the nonlinear association between PLR and survival is another strength of this study.
However, our study had several limitations. First, its single-center, retrospective design may have caused potential bias. The systemic inflammatory marker cutoffs were derived from the AUC of these parameters against overall survival. Moreover, these cutoffs were not assessed in another data set for validation. Therefore, to use PLR in clinical practice, further prospective study or consensus on the optimal cutoff would be desirable. Secondly, our study could not set appropriate criteria for application, considering the major limitation of the systemic inflammatory markers, i.e., the influence of infection. Accumulation of scientific evidence and an understanding of the role of systemic markers through prospective multicenter trials would be needed to address this limitation.
In conclusion, this study evaluated the prognostic value of a panel of systemic inflammatory markers, including NLR, PLR, SII, and SIRI, in OSCC. Our data demonstrated that PLR was a valuable independent prognostic biomarker in patients with OSCC, especially in those with advanced disease. High PLR was associated with worse survival in these patients. We also demonstrated a nonlinear correlation between PLR and survival. A Jshaped association implied that an extremely low PLR was also a poor prognostic factor in patients with OSCC. Preoperative assessments of cellular biomarkers from peripheral blood could provide high-quality prognostic information, and they represent another promising approach for improving patient stratification.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biomedicines10061268/s1, Figure S1: Pearson's correlation between lymphocyte count and platelet count, neutrophil count, and monocyte count; Figure S2: Pearson's correlation between platelet-lymphocyte ratio and neutrophil-lymphocyte ratio; Table S1: Correlation between clinicopathologic parameters and white blood cell and platelet count.

Informed Consent Statement:
The requirement for informed consent was waived by the Institutional Review Board.