The role of systemic inflammatory response index (SIRI) and tumor-infiltrating lymphocytes (TILs) in the prognosis of patients with laryngeal squamous cell carcinoma

Systemic inflammatory response index (SIRI) values and tumor-infiltrating lymphocytes (TILs) are associated with the prognosis of various tumors. There is minimal evidence of those two as prognostic markers in laryngeal squamous cell carcinoma (LSCC). In this study, we aimed to examine the predictive value of SIRI and tumor-infiltrating CD3+/CD4+/CD8+ T cells in the prognosis of patients who underwent partial or total laryngectomy. A total of 78 patients with LSCC who underwent total or partial laryngectomy at the Eye, Ear, Nose, and Throat Hospital of Fudan University between 2013 and 2015 were retrospectively analyzed. The tumor tissues of 78 LSCC patients were retrospectively evaluated using immunohistochemical staining for CD3+ /CD4+ /CD8+ -cells. The overall survival (OS) and disease-free survival (DFS) rates were recorded using the Kaplan–Meier method. Patients with high immunoscore (IS) (3–4) had prolonged survival (P < 0.001 for OS). High SIRI values were independently associated with poorer OS and DFS (P = 0.018 for OS; P = 0.016 for DFS). CD8+ TILs and SIRI values showed a- negative association (P < 0.01). Patients with low SIRI values and high IS had better 5-year OS and DFS than those with high SIRI values and low IS (P < 0.001 for OS; P = 0.0014 for DFS). Patients with ‘hot’ tumor had a higher 5-year OS than those with ‘excluded’ or ‘cold’ phenotype. The SIRI values and the density of TILs may help predict LSCC patients' outcomes after surgery. The combination of SIRI and IS may be a new component of the tumor, nodes, and metastases (TNM) classification of cancer and prognostic factor for T-cell-target immunotherapy.


Introduction
In 2021, there were 12,620 new cases of laryngeal cancer and 3,770 deaths in the United States (Siegel et al. 2021). The incidence of laryngeal cancer has decreased by 2.4% per year over the past 10 years. However, the 5-year survival rate for laryngeal cancer remained at approximately 60.9%. Among patients with advanced laryngeal cancer (Stage III and IV), the survival rate declined to 40%, which has remained virtually unchanged over the past 10 years (Nachalon et al. 2020;Lahane et al. 2020). Current studies on the prognostic factors associated with LSCC patients have primarily focused on age, TNM, and treatment options (Nachalon et al. 2020). Laryngeal cancer requires complex treatment strategies and long-term follow-up, which impose a severe financial burden on patients and significantly reduce Tian Wang and  1 3 their quality of life (Sharpe et al. 2019). Therefore, it is essential to identify the prognostic parameters that accurately estimate the treatment outcome. For the prognostic indicator to be widely used, it must be generally evaluated, readily available, reusable, low cost, and easily standardized. Peripheral blood meets all these criteria. Qi and colleagues recently proposed SIRI, a parameter based on neutrophils, monocytes, and lymphocytes counts in peripheral blood before receiving chemotherapy (Qi et al. 2016). SIRI is defined as (N*M)/L, where N, M, and L were the absolute count of neutrophils, monocytes, and lymphocytes obtained from the pretreatment blood sample. Valero C and colleagues showed that patients with higher SIRI have a significantly reduced disease-specific survival (DSS) in the head and neck squamous cell carcinomas (HNSCC) (Valero et al. 2020). However, SIRI can only evaluate the systemic immune status, and it is impossible to assess the tumor immune microenvironment.
Inflammation and immune escape have been identified as two significant cancer features (Hanahan and Weinberg 2011), and are both contributed to the progress of cancer and worse survival outcomes in patients (Diakos et al. 2014). TILs are present in the tumor lesions and the interstitial space. The type, density, and location of TILs are of great value in assessing the prognosis of patients with malignant tumors, especially inflammation-related tumors (Barua et al. 2018;Fridman et al. 2012). The immune scoring system summarizes the density of CD3+ and CD8+ T-cell effectors inside the tumor and within its infiltrative margins. The IS combined with the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/ UICC) staging has been shown to determine the prognosis of cancer patients (Galon et al. 2014). The prognostic value of IS in colorectal, skin, breast, and other cancers has been widely studied. Several studies have demonstrated its association with the overall survival (OS) and disease-free survival (DFS) (Galon et al. 2014;Zeng et al. 2021;Sun et al. 2019). Like other tumors, LSCC was closely associated with inflammation, and the tumor is infiltrated heavily by a complex array of inflammatory and lymph-like cells. As a concept that quantifies the number, type, and location of TILs, the IS has been speculated to be potentially used as a prognostic factor in the evaluation of LSCC (Zhang et al. 2018).
SIRI is measurable and can act as an indicator of immune status in the peripheral blood in response to inflammation. In contrast, TILs detected by immunohistochemical staining can be considered a reflection of immune cell infiltration within tumor lesions. However, no study has been conducted with both the SIRI and TILs that analyzes their expected values for the prognosis of LSCC. Little evidence exists on the mechanisms linking poor prognosis in cancer patients to systemic inflammation. We designed this study to investigate further the predictive value of SIRI and the density of TILs in the prognosis of LSCC.

Patients and study design
A total of 78 patients with primary LSCC between 2013 and 2015 were included in this study. The inclusion criteria were as follows: (1) histopathological confirmation of primary LSCC; (2) absence of pretreatment; (3) availability of at least one block of formalin-fixed and paraffin-embedded tissue for tissue microarray (TMA); (4) absence of distant metastases or second cancer. These patients were treated with total or partial laryngectomy at the Otolaryngology-Head and Neck Surgery Department of the Eye, Ear, Nose, and Throat Hospital. None of the patients had adjuvant radiotherapy after surgery. According to the AJCC, all patients were staged using resected specimens. Required parameters of peripheral blood were the absolute count of neutrophils (N), mononuclear cells (M), platelet (P), and lymphocyte (L). The Institutional Ethics Committee approved this study at the Eye, Ear, Nose, and Throat Hospital of Fudan University. All participants gave informed consent to take part in the study.

Immunohistochemical evaluation
Sections of 4-5 μm thick were cut from each selected block of formalin-fixed quasi-embryonic TMA tissue. It was followed by de-paraffinization with xylene for 10 min and rehydration through a series of graded alcohols (100%, 90%, 85%, 80%, 75%, 60%, 50%, 30%), at each concentration for 5 min. The antigen was retrieved via heating for 15 min in a microwave oven. The TMA was placed in triple-buffered saline (TBS) Tween buffer for 6 min when cooled to room temperature. Subsequently, the slides were blocked and labeled with respective primary antibody. TILs were quantified on TMA. The stained TMA sections were imaged using an Aperio digital slide scanner (Leica Biosystems) and then analyzed using QuPath v. 0.2.3 (Queen's University, Belfast, Northern Ireland). TILs were calculated as the sum of individual cells from the tumor interior (TI) and the tumor stroma (TS). The median number of TILs was used to classify patients into low-infiltrating or high-infiltrating groups (CD3+ , CD8+ , CD4+ , and Foxp3 + expression).
Based on the infiltration of CD3+ and CD8+ T cells, the mean of the four percentiles (two markers, two regions) was calculated and converted into an IS. Specifically, 1) IS = 0, the patient had low infiltration of both CD3+ and CD8+ ; 2) IS = 1, a high density of at least one marker in one region (TI or TS); 3) IS = 2-4, the total score for each region of high density   (Fig. 1C).

Statistical analysis
The OS was calculated from the start of treatment to the date of death. The DFS was defined as the time from treatment initiation to tumor recurrence (local area recurrence and/or distant metastases) or death from non-tumor-related causes. The OS and DFS curves were plotted according to the Kaplan-Meier method using SPSS (version 25.0) and were generated with GraphPad Prism (version 8.0). The univariate and multivariate analyses for the risk of OS and DFS were performed using the log-rank (Mantel-Cox) test and the Cox proportional hazard models, respectively. Associations between TILs density, SIRI, and recorded clinicopathological characteristics were assessed using χ 2 tests for categorical variables. A P value of 0.05 (two-tailed) was considered statistically significant.

Characteristics of LSCC patients
The patients' baseline clinical characteristics are described in Table 1. The median age of the patient population was 63.5 years (range from 47 to 84 years), and most of the patients were male (77, 98.7%). 67.9% of patients had a smoking history, and 35.9% had a drinking history. There were no patients in the T1 stage, but 19 patients (24.4%) were in the T2 stage, 42 patients (53.8%) were in the T3 stage, and 17 patients (21.8%) were in T4. 61.5% of patients did not have pathological nodal involvement, while the others had (10 for N1, 20 for N2). More than half of the patients (65, 77.4%) were diagnosed with stage III or IV LSCC. This study included 38 (45.2%) patients with supraglottic carcinoma, 30 (35.7%) with glottic carcinoma, 5 (6%) with subglottic carcinoma, and 5 (6%) with transglottic carcinoma. Using the Kaplan-Meier methodology, the estimated probability of DFS was found to be 79.1% at 1 year, 72% at 3 years, and 55% at 5 years, and the estimated QuPath v.0.2.3 to define TI and TS in the scanned diagram and calculate the CD3+ and CD8+ values in them, respectively probability of OS was 91% at 1 year, 67.7% at 3 years, and 49.4% at 5 years.
pT stage was significantly associated with 5-year OS but not with DFS (P = 0.029 for OS and P = 0.07 for DFS). Lymph-node involvement (N+) predicts poor OS(P = 0.034) and DFS(P = 0.049). Patients with tumor size less than 3 cm were associated with better 5-year OS compared to those with tumor size greater than 3 cm (P = 0.046). Alcohol consumption was associated with poorer OS (P = 0.042). No effect of age, smoking, and other clinical indexes on survival prognosis was observed (P > 0.05) (Supplementary Table 1 and Supplementary Fig. 1).

Association of TILs' density and SIRI with clinical features
Immunohistochemical analysis revealed that the sum of CD3+ expression in TS and TI was defined as 'low' (< 840) in 39 patients (50%) based on the median score. Similarly, the sum of CD8+ expression in TS and TI was defined as 'low' (< 198) in 39 patients (50%) based on the median score. Meanwhile, 74 patients (94.8%) were defined as having 'low' CD3+ expression in TS, while 46 patients (58.9%) were defined as having 'low' CD3+ expression in TI. A total of 65 patients (83.3%) were defined as having 'low' expression of CD8+ in TS, while 36 patients (46.1%) were defined as having 'low' expression of CD8+ in TI. The IS calculated via QuPath v.0.2.3 ( Fig. 1) showed that 38 patients (48.7%) had a lower score (0-1), while 40 patients (51.2%) had a higher score (2-4). Based on the median score, the SIRI value was defined as 'low' (< 1.21) in 39 patients (50%). In comparison to patients with N0 stage, those with advanced N stage (N1-2) were significantly associated with increased levels of Foxp3 + and CD4+ in TS and high SIRI values, indicating a more aggressive disease ( Table 2).

Association of TILs and SIRI with survival outcomes
In our study, Kaplan-Meier analysis showed that lower CD8 + TILs (in either TI or TS) were correlated to the worse 5-year OS (P = 0.037 for TI; P = 0.041 for TS) but not significantly correlated to the 5-year DFS (P = 0.369 for TI; P = 0.89 for TS). Patients with a high density of CD3 + TILs in TS showed improved 5-year OS and DFS compared to those with a low density of CD3 + TILs (P < 0.001 for 5-year OS; P = 0.025 for 5-year DFS) ( Supplementary  Fig. 2). Patients with a high density of both CD3 + and CD8 + (combined region of TS and TI) were associated with better 5-year OS and DFS (P < 0.001 for both OS; P = 0.038 for CD3+ , DFS; P = 0.023 for CD8 + , DFS). Patients with a high value of SIRI showed poorer 5-year OS and DFS compared to those with a low value of SIRI (P = 0.014 for 5-year OS; P = 0.012 for 5-year DFS) (Fig. 2). No significant correlation was observed between the density of CD4 + or FOXP3 + and the patients' survival (Supplementary Table 1). Patients were classified according to the IS scores, including IS0, IS1, IS2, IS3, and IS4. There were 33.3%, 15.4%, 16.7%, 14.1%, and 20.5% of patients in the respective subgroups. Patients with high IS (3-4) had the longest survival and the lowest risk of recurrence (P < 0.001 for OS, P = 0.007 for DFS) (Fig. 2).
As shown in Table 3, increased CD8 + TILs' infiltration and lower IS were associated with higher SIRI (P < 0.05). Patients with low peripheral blood SIRI values and high IS had a higher 5-year OS and DFS (65.7%, P < 0.001; 80.7%, P = 0.01). However, patients had a lower 5-year OS and DFS when having high SIRI and low IS (25%, P < 0.001; 36.7%, P = 0.01). Patients had better 5-year OS and DFS when having high values on both peripheral blood SIRI and IS compared to those with low values on both peripheral blood According to CD3 + /CD8 + T cell infiltrated, the terms 'hot' (IS4, highly infiltrated), 'excluded' (IS 1-3, highly infiltrated in TS and non-infiltrated in TI), and 'cold' (IS0, non-infiltrated) were used to classified tumor based on T-cell landscape rather than tumor itself. Patients with 'hot' tumor had a higher 5-year OS than those with 'excluded' or 'cold' phenotype (Fig. 3). Compared with 'cold' phenotype, 'hot' or 'excluded' phenotype had higher CD4 + T-cell infiltration and lower SIRI.
In the univariate analysis, we found that lymph-node involvement, alcohol consumption, a lower density of TILs (CD3 + , CD8 + ), high SIRI, and a lower IS (0-1) were significantly associated with the OS of patients (P < 0.05). Meanwhile, high SIRI, a lower density of TILs (CD3 + , CD8 + ), and IS (0-1) were significantly associated with the DFS of patients (P < 0.05). In the multivariate analysis, we found that the immune score and lymph-node involvement (N+) were significantly associated with OS (P < 0.05). The SIRI stratification was significantly associated with patients' DFS (P < 0.05) (Table 4).

Discussion
The incidence of laryngeal cancer has decreased each year for the past decades. However, the 5-year survival rate has not improved, especially for patients with intermediate and advanced stages, which is only 40% (Yang et al. 2019). Due to the specific site of laryngeal cancer and individual differences in sensitivity to radiotherapy, the development of treatment plans is often complicated. After the treatment, long-term follow-up is required, which imposes a more significant financial burden on the patients. The TNM system is commonly used to stage laryngeal cancer and aid in prognosis, but some patients within the same clinical stage can show notably different tumor outcomes. This suggests that inflammation plays a vital role in the development of tumors. By orchestrating the tumor microenvironment, inflammatory cells influence the proliferation and migration of tumor cells (Chao et al. 2020). Our study found that the SIRI values (systemic inflammation) and the density of TILs (locoregional inflammation) may help predict LSCC patients' outcomes after surgery. CD3 + T lymphocytes are the primary immune cells in the tumor microenvironment. Previous studies have reported the association of high CD3 + expression with better prognosis in pancreatic, rectal, head and neck squamous, and laryngeal cancers (Miksch et al. 2019;Zhou et al. 2019;Nguyen et al. 2016). Our study suggested that high CD3 + expression in TI and TS was associated with a better prognosis. CD8 + T lymphocytes, also known as cytotoxic T cells, bind to human leukocyte antigen I, directly target and destroy tumor cells. A meta-analysis found that a high density of CD8 + T lymphocytes was associated with a good prognosis for gastric cancer (Lee et al. 2018). In our study, high CD8 + expression in either TI or TS was associated with better 5-year OS but not 5-year DFS. Furthermore, the univariate analysis showed that the high CD8 + expression in TI and TS was independently correlated with OS and DFS. The IS reflects the infiltration of CD3 + and CD8 + in TI and TS. It is highly reproducible, comprehensive, and widely used in the study of colorectal cancer (Pages et al. 2018;Galon and Lanzi 2020). Zhang and colleagues demonstrated that the LSCC patients with IS 0-1 had worse OS and DFS than those with IS 2-4, which is in line with our findings in previous study. (Zhang et al. 2021). The application of IS (introduced by Galon et al. 2012) for predicting prognosis and drug sensitivity in colorectal cancer patients has been widely studied. Some studies suggested that the IS is a better approach than the current TNM staging system for stage I-III colorectal  Lea et al. 2014). Currently, the terms 'hot' and 'cold' are widely used to classify tumor based on immune contexture rather than clinical-pathological feature. The 'excluded' phenotype refers to the tumor immune escape by hindering T-cell infiltration (Galon and Bruni 2019). In our study, patients with 'hot' tumor had a higher 5-year OS than those with 'excluded' or 'cold' phenotype, which supports that the tumor classification based on immune contexture could be a good prognostic factor.
In this study, we found that high SIRI values were independently associated with poorer OS and DFS, which agrees with the previous reports (Valero et al. 2020;Wang et al. 2020a;Sun et al. 2020;Geng et al. 2018). Unlike other molecular diagnostics, SIRI is a composite inflammatory marker formed by neutrophil, monocyte, and lymphocyte counts. It is rapid, inexpensive and easily obtained from routine blood tests. Based on the three types of inflammatory cells (Neutrophils, Macrophages, Lymphocytes), SIRI can be used to measure the balance between the pre-tumor inflammatory state and the anti-tumor immune state in patients with tumors (Chen et al. 2019). Neutrophils can suppress the immune system by releasing cytokines and chemokines, promoting circulating tumor cells, and stimulating tumor angiogenesis and progression (Kolaczkowska and Kubes 2013;Tan et al. 2013). The density of tumor-associated macrophages has been shown to affect tumor angiogenesis and is associated with poor prognosis (Mantovani et al. 2006). As a result, a higher SIRI value may reflect an imbalance between the host inflammatory response and the immune response, biased toward a cancer-promoting effect. TILs represent the locoregional immune microenvironment around tumor tissue. The IS reflects a combined CD3+ and CD8+ infiltration in TI and TS, which is increasingly relevant for different cancer types, especially inflammation-related tumors.
For the first time in this study, the integration of SIRI and IS was used to assess tumors' prognosis. We supposed that systemic inflammatory variation might be related to tumor microenvironment change. To verify this hypothesis, we analyzed the correlation between SIRI and TILs (CD8+) and SIRI are negatively correlated, which means the ratio of lymphocytes in peripheral blood impact TILs around tumor microenvironment. High IS and low SIRI group had the longest survival and lowest risk of recurrence at 5 years, compared with other groups. Although the prognostic value of SIRI was widely demonstrated, the value of SIRI was influenced by various factors including acute inflammation, drug and autoimmune disease. The combination of SIRI and IS would better show anti-tumor immune status of patients with LSCC.
Patients with lymph-node metastasis have a worse prognosis in LSCC. The location of the tumor and T staging are considered the independent risk factors for lymph-node metastasis in laryngeal cancer (Wang et al. 2020b;Chen et al. 2021). A recent meta-analysis of 2835 gastric cancer patients from 9 studies suggested a correlation between TIL infiltration and lymph-node involvement (Tian et al. 2021). In our study, high peripheral blood SIRI values were found to be correlated with a high incidence of lymph-node metastases. Increased expression of Foxp3+ and CD4+ in the TS was associated with a higher rate of lymph-node involvement. Foxp3+ is an essential immunosuppressive cell that can help tumor cells to achieve immune escape by interfering with the antigen presentation of tumor cells and inhibiting the T-cell killing function. Foxp3+ may indirectly inhibit tumor-specific T cells by releasing factors, such as TGF-β and IL-10, exerting immunosuppressive functions and promoting immune escape of tumors (Goto et al. 2020). The role of CD4+ T lymphocytes in the tumor microenvironment is unclear, and therefore, the prognostic value of CD4+ T lymphocytes in tumors is somewhat controversial. In our study, high CD4+ T expression is associated with increased rate of lymph-node metastases, acting as a promoter of immune escape.  The patients were then classified as 'hot' (IS4, highly infiltrated), 'excluded' (highly infiltrated in TS and non-infiltrated in TI), and 'cold' (IS0, non-infiltrated) phenotype. The 'hot' tumor had the longest survival (C). Compared with 'cold' phenotype, 'hot' or 'excluded' had more CD4+ TILs and lower SIRI