1. Introduction
Oral squamous cell carcinoma (OSCC) is the most common tumor in the oral and maxillofacial region, accounting for 90% of lip and oral cavity cancer [
1] and ranking as the 16th most common malignancy worldwide [
2]. The global annual estimated incidence of lip and oral cavity cancer was 389,485 incident cases in 2022 [
2]. According to the National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines Version 1.2022 [
3], the mainstream management of OSCC remains radical surgery combined with sequential therapy. This approach, however, compromises patients’ quality of life due to the extensive excision and neck dissection, followed by postoperative radiation. Such treatments can impair critical functions of the oral and maxillofacial region, including eating, drinking, swallowing, and speaking, as well as alter appearance, posing significant threats to patients’ lives. In most countries, the five-year survival rates for cancers of the tongue, oral cavity, and oropharynx are around 50% [
4], with a total of 188,230 deaths from cancer of the lip and oral cavity reported in 2022 [
2]. Therefore, it is urgent to identify the most effective treatment strategy for the appropriate patients.
Although various neoadjuvant therapies have been tested in clinical trials, challenging the traditional standard therapy (radical surgery and postoperative radiation), no single treatment strategy has been universally beneficial. In our previous phase III clinical trial, docetaxel, cisplatin, and 5-fluorouracil (5-FU) (TPF) were used as induction chemotherapy in patients with locally advanced oral squamous cell carcinoma (LAOSCC) (registration ID: NCT01542931). However, only a portion of pathologic responders benefited from this approach [
5]. Other phase III clinical trials have shown varying survival benefits from additional TPF induction chemotherapy [
6,
7,
8]. In the NCT00095875 trial, TPF induction chemotherapy did not provide extra survival benefits beyond concurrent chemoradiotherapy in locally advanced head and neck cancer [
6]. Conversely, the NCT01245959 trial demonstrated that adding TPF induction chemotherapy to concurrent chemoradiotherapy significantly improved failure-free survival in locoregionally advanced nasopharyngeal carcinoma with acceptable toxicity [
7]. Similarly, the NCT01086826 trial showed that TPF induction chemotherapy significantly improved radiological complete responses, progression-free survival, and overall survival without compromising compliance with concomitant platinum-based chemoradiotherapy in locally advanced head and neck squamous cell carcinoma [
8]. These findings highlight the potential value of neoadjuvant therapy, suggesting that only a subset of patients benefit from TPF induction chemotherapy [
5]. This underscores the need for a personalized treatment strategy to identify specific patients who will benefit. The rise of personalized medicine and the increasing importance of biomarkers in tailoring treatment strategies have led to numerous clinical trials exploring different approaches. In our previous studies, growth differentiation factor 15 (GDF15) was identified as a potential predictive biomarker, with patients exhibiting cN- and high expression of GDF15 benefiting from TPF induction chemotherapy in LAOSCC [
9]. Additionally, normal body mass index (BMI) was found to predict survival benefits from TPF induction chemotherapy in patients with stage IVA cancer within the same cohort [
10].
The derived neutrophil to lymphocyte ratio (dNLR) is an index easily obtained from a routine complete blood count (CBC) test. It is calculated as the ratio of neutrophils to the difference between leukocytes and neutrophils, reflecting the relative quantity of tumor-immune-related cells in leukocytes. A smaller dNLR indicates more tumor-immune-related cell infiltration [
11]. The prognostic value of the dNLR has been demonstrated in various cancers, including pancreatic ductal adenocarcinoma [
12], breast cancer [
13], non-small cell lung cancer [
14,
15], advanced or metastatic melanoma [
16,
17], metastatic prostate cancer [
18,
19], metastatic renal cell carcinoma [
20], advanced or metastatic colorectal cancer [
21,
22], intrahepatic cholangiocarcinoma [
15], and upper tract urothelial carcinoma [
23]. However, in the field of head and neck squamous cell carcinoma (HNSCC), there is no clear evidence supporting the dNLR as a prognostic biomarker [
24,
25]. It has been suggested that the slightly decreased prognostic value might be due to a smaller area under the curve-receiver operating characteristic (AUC-ROC) for predicting overall survival (OS) [
24]. Additionally, no literature has reported on the relationship between the dNLR and survival benefits from TPF induction chemotherapy. Despite this, TPF induction chemotherapy has shown benefits in previous trials (NCT01542931, NCT01245959, and NCT01086826) [
5,
7,
8]. Therefore, it is crucial to identify the exact patient population that will benefit from TPF induction chemotherapy. In summary, this study aimed to demonstrate our findings on the predictive value of the dNLR for survival benefits from TPF induction chemotherapy in patients with LAOSCC.
4. Discussion
In this study, the CBC, clinical characteristics, and survival outcomes of LAOSCC patients were retrospectively analyzed. In the control group, through univariate and multivariate Cox model analyses, the baseline dNLR (as a continuous variable) was confirmed as an independent prognostic factor for survival outcomes (
Table 3 and
Table 4). Using the ROC curve and Youden’s index (
Figure 1), a dNLR cutoff point of 1.555 was established for all survival outcomes (OS, DFS, LRFS, and DMFS). Based on this cutoff point, patients were categorized into two subgroups: low dNLR and high dNLR. Patients with a low baseline dNLR had better survival outcomes (
Figure 2), and this finding was consistent in the experimental group (
Table 5 and
Table 6). Concurrently, the cTNM stage (as a categorical variable) was confirmed as an independent prognostic factor for survival outcomes. When combining the two variables—cTNM stage III disease and low dNLR (dNLR ≤ 1.555)—this specific patient population showed significant benefit from TPF induction chemotherapy (
Figure 3).
In recent years, many immune biomarkers easily acquired from a CBC have been introduced into the field of tumor therapy as prognostic factors. These include the neutrophil to lymphocyte ratio (NLR) [
28,
29,
30], the derived neutrophil to lymphocyte ratio (dNLR) [
11,
14,
18,
31], the platelet to lymphocyte ratio (PLR) [
32,
33,
34], the lymphocyte to monocyte ratio (LMR) [
20,
35], and the pan-immune inflammation value (PIV) [
36,
37]. In OSCC, the NLR has been reported as a significant independent predictor of disease-specific survival (DSS) [
38,
39] and is significantly correlated with stromal infiltration of CD8+, CD4+, and CD20+ lymphocytes [
40]. The dNLR has been correlated with the occurrence of complications [
41]. The PLR has shown a stronger association with DSS and progression-free survival (PFS) in patients who are male, have stage III/IV OSCC, or have lymph node metastasis [
42]. The NLR and dNLR appear similar and have comparable effects on cancer-specific mortality [
43], and a positive correlation between the dNLR and NLR has been found [
44]. However, the calculation method for the dNLR includes not only lymphocytes but also monocytes and other subtypes of immune-related cells. In this study, dNLR, NLR, PLR, LMR, and PIV were analyzed using univariate and multivariate Cox regression analyses. Only the dNLR demonstrated an independent negative effect on survival prognosis (
Table 3,
Table 4,
Table 5 and
Table 6). We believe the complexity of the tumor immune microenvironment and the neutrophil heterogeneity among OSCC patients contribute to these results.
To our knowledge, this is the first time that the dNLR has been confirmed as an independent prognostic factor for the survival of OSCC patients through multivariate analysis. These patients were treated either with surgery and postoperative radiation or with TPF induction chemotherapy, surgery, and postoperative radiation. Additionally, this is the first time that the dNLR level and cTNM stage have been confirmed as criteria for identifying patient populations that can benefit from TPF induction chemotherapy (
Figure 3). In our previous phase III clinical trial (registration ID: NCT01542931), TPF induction chemotherapy did not improve survival compared with upfront surgery [
5]. However, patients with cN2 disease appeared to have improved OS (HR, 0.418; 95% CI, 0.179 to 0.974;
p = 0.043) and DMFS (HR, 0.418; 95% CI, 0.179 to 0.974;
p = 0.043) when treated with TPF induction chemotherapy compared to those who were not [
5]. Thus, efforts were made to identify clinically valuable biomarkers to screen appropriate OSCC patient populations for TPF induction chemotherapy. Lymph node ratio (LNR), the ratio of pathologically confirmed positive lymph nodes to the total number of surgically removed lymph nodes, was found to be connected to prognosis and could be an independent prognostic factor. OSCC patients with high-risk LNR (>7.6%) or positive extranodal extension (ENE) had significantly worse clinical outcomes than patients with low-risk LNR (≤7.6%) or negative ENE [
45]. In another of our previous studies, BMI was also found to be an independent prognostic factor. Compared to normoweight patients, overweight and obese patients had better clinical outcomes, while underweight status was associated with poor survival [
10]. Furthermore, normoweight patients with cTNM stage IVA disease benefited from TPF induction chemotherapy followed by surgery and postoperative radiation, compared to surgery and postoperative radiation alone, in terms of OS and DMFS. In this study, the dNLR was found to be an independent prognostic factor for assessing the survival of OSCC patients, even after adjusting for other important variables. This finding is consistent with many other clinical trials that reported a significant relationship between low dNLR values and a good prognosis or favorable clinical or pathological responses [
11,
14,
18,
31,
43,
46,
47].
In the present study, to account for the potential influence of TPF induction chemotherapy on inflammation biomarkers, hematology and survival data from the control group were used to form the ROC curve and calculate the AUC for the dNLR cutoff point. All four survival rates pointed to the same dNLR cutoff point of 1.555. This demonstrated that the dNLR can be an independent prognostic factor for the survival of OSCC patients based on multivariate analysis. There have been other reported dNLR cutoff points for different types of tumors. For instance, a dNLR cutoff point of 1.775 was found useful in predicting metastatic disease in testicular germ-cell tumors [
47]. Another study on breast cancer reported a baseline dNLR cutoff point of less than 1.715 in predicting pathological complete response [
46]. In patients treated with the immune checkpoint inhibitor pembrolizumab for non-small cell lung cancer, those with a dNLR cutoff < 2.6 had significantly higher objective response rates (ORRs), longer median progression-free survival, and higher numbers of tumor-associated CD8+, FOXP3+, PD-1+ immune cells, and PD-1+ CD8+ T cells [
11]. In the Lung Immune Prognostic Index (LIPI), a dNLR cutoff of 3 was used to determine the prognosis of non-small cell carcinoma [
14]. Various factors, such as tumor type, TNM stage, systemic inflammation, and population differences, contribute to these differing cutoff points. This may indicate the more malignant biological behavior of OSCC compared to other tumor types.
It was reported that a lower baseline dNLR level is associated with a higher number of tumor-associated immune cells [
11]. In this study, a statistically significant correlation between the dNLR and cTNM stage (
Table 1) was found in the control group, suggesting a mutual relationship where cTNM stage IVA disease often coincides with a high dNLR. Similar results have been reported for colorectal cancer, renal cell carcinoma, and gastric cancer [
48,
49,
50], where TNM stage is positively related to dNLR. A low dNLR is associated with significantly increased tumor-associated CD8+, FOXP3+, and PD-1+ immune cells and favorable outcomes [
11]. Stage III and stage IVA are all considered local advanced diseases in OSCC; thus, this difference was neglected in the following analysis.
A high baseline dNLR level indicates a relatively higher number of neutrophils. Neutrophils exhibit diversity, heterogeneity, and plasticity in cancer [
51,
52,
53]. Different subgroups of neutrophils, each involved in distinct biological processes, have been reported [
52,
54]. In mouse and human lung cancer models, seven populations of neutrophils were identified, and CD40 agonist antibody treatment increased immune response by more than 10-fold in both N1a (Sellhi Ngphi) and N2 (Sellhi Cxcl10hi) neutrophil populations, characterized by high expression of interferon-stimulated genes (ISGs) [
54]. This suggests that individuals with high baseline dNLR levels may lack the capacity for neutrophil differentiation into ISG-high expression subtypes, leading to an unfavorable prognosis. Furthermore, in the tumor immune microenvironment, tumors can modulate neutrophil extracellular traps (NETosis), causing NET-associated complications like thrombosis. NETosis captures tumors, promotes their growth, and leads to subsequent metastasis [
55]. In breast cancer, the tumor-secreted protease cathepsin C promotes breast-to-lung metastasis by regulating the recruitment of neutrophils and the formation of neutrophil extracellular traps [
56].
A high baseline dNLR level can also be due to low levels of leukocytes other than neutrophils, primarily lymphocytes. Lymphocytes play a crucial role in tumor immunity. However, the functional state of T lymphocytes is also critical to tumor immunity. T cell dysfunction in human cancer is associated with changes in T cell functionality rather than inactivity [
57]. Although low dNLR levels predict a better prognosis in this study, the state of lymphocytes should still be considered.
There were some limitations to this study. Typically, patients with LAOSCC diagnosed as cTNM stage IVA have worse survival outcomes, which could lead to potential bias. Fortunately, this was not observed in the TPF induction chemotherapy group. This issue could be addressed in another clinical trial with a larger sample size to verify the possible connection between TNM stage and baseline dNLR in LAOSCC. Additionally, the retrospective nature of the study may include unnecessary confounding factors, and another clinical trial is needed to confirm whether the dNLR cutoff point is a suitable boundary for other survival rates or OSCC patients with different systemic conditions or TNM stages.