Development and validation of immune inflammation–based index for predicting the clinical outcome in patients with nasopharyngeal carcinoma

Abstract Inflammation indicators, such as systemic inflammation response index (SIRI), systemic immune‐inflammation index (SII), neutrophil‐to‐lymphocyte ratio (NLR) and platelet‐lymphocyte ratio (PLR), are associated with poor prognosis in various solid cancers. In this study, we investigated the predictive value of these inflammation indicators in nasopharyngeal carcinoma (NPC). This retrospective study involved 559 patients with NPC and 500 patients with chronic rhinitis, and 255 NPC patients were followed up successfully. Continuous variables and qualitative variables were measured by t test and chi‐square test, respectively. The optimal cut‐off values of various inflammation indicators were determined by receiver operating characteristic (ROC) curve. Moreover, the diagnostic value for NPC was decided by the area under the curves (AUCs). The Kaplan‐Meier methods and the log‐rank test were used to analyse overall survival (OS) and disease‐free survival (DFS). The independent prognostic risk factors for survival and influencing factors of side effects after treatment were analysed by Cox and logistic regression analysis, respectively. Most haematological indexes of NPC and rhinitis were significantly different between the two groups, and PLR was optimal predictive indicators of diagnosis. In the multivariable Cox regression analysis, PLR, WBC, RDW, M stage and age were independent prognostic risk factors. Many inflammation indicators that affected various side effects were evaluated by logistic regression analysis. In conclusion, the combined inflammation indicators were superior to single haematological indicator in the diagnosis and prognosis of NPC. These inflammation indicators can be used to supply the current evaluation system of the TNM staging system to help predict the prognosis in NPC patients.


| INTRODUC TI ON
Nasopharyngeal carcinoma (NPC) is a malignant epithelial cancer that occurs in the epithelial lining of the nasopharynx with the highest rate of metastasis among head and neck cancers. 1 NPC has an extraordinarily skewed geographic distribution worldwide, which is mainly prevalent in southern China and South-East Asian countries. 1 More than 129 000 new cases of NPC were reported worldwide, and the incidence of the male is higher than that of female. 1 The mortality from cancer is mostly attributable to metastases, not the primary cancers, and the effective treatment for cancer depends mainly on our capacity to reverse the process of metastasis. 2 Intensity-modulated radiation therapy (IMRT) and concurrent chemotherapy are regarded as the primary treatment for NPC. 3 However, the treatment is related to acute and late toxicities with impairment of patients' quality of life, 4 such as dysphagia. 5,6 Other side effects, such as the arrest of bone marrow, radiation stomatitis and dermatitis, need to be further investigated.
The classification method of NPC is mainly relied on the tumour-node-metastasis (TNM) staging criteria, which is used for treatment selection, cancer control activities and outcome prediction. However, the failure to consider the functional status of NPC leads to different prognoses in patients with the same TNM staging. 7 More reliable markers are necessary to supply clinical diagnosis and treatment.
The inflammatory responses play an essential role in various stages of cancer development, including occurrence, progression, malignant conversion, invasion and metastasis, and moreover, the inflammation affects immune surveillance and responses to therapy. 8 Solid malignancies trigger an intrinsic inflammatory response and then building up a pro-tumorigenic microenvironment, which promotes the development of cancers. 9 Cancers contain various noncancerous cells including immune cells, such as T cells, macrophages and neutrophils. These cells can be anti-or tumorigenic and associate with survival in several cancer types. 10 The inflammation indicators including neutrophils, 11 lymphocytes and monocytes, 12 and red cell volume distribution width (RDW) 13 have prognostic value in cancers. The integration of two types of white blood cell indicators, such as the neutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR) and lymphocyte-monocyte ratio (LMR), is considered to be independent prognostic factors for colorectal cancer. 14 15,16 These inflammation indexes are also considered to be independent prognostic factors for cancers, and their prognostic value is higher than that of only white blood cells.
However, the cut-off value of immune-inflammation indicators is diverse in different cancers. The cut-off value of SII, NLR and PLR in non-small-cell lung cancer is 660, 3.57 and 147, respectively, 16 while these values in metastatic prostate cancer are 535, 3 and 210, respectively. 17 There are few reports on the relationship between combined inflammation indicators and NPC prognosis, and the basophil has never been reported in NPC prognosis.
In this study, we investigated the efficiency of these inflammation indicators on the diagnosis of NPC, and these inflammation indicators can be established as a mechanism for predicting prognosis of cancer patients in clinical settings that would help for future novel treatments.

| Patients
We retrospectively recruited 559 patients who were diagnosed as

| Inclusion and exclusion criteria
The inclusion criteria in this study comprised of: (a) patients with histopathological confirmation of NPC; (b) patients with proper renal, cardiac and liver function to tolerate chemotherapy and radiotherapy; and (c) patients with a complete record of haematological indicators. Exclusion criteria were as follows: (a) patients with other types of malignancy. Finally, we have retrieved data of 255 patients with complete follow-up data using for survival analysis.

| Haematological examination
Fasting whole blood from every patient was collected in an EDTA anticoagulant-treated tube on the admission without the next step of treatment, and analysed within 30 minutes of collection.

| Follow-up
We chose the OS and DFS as the primary end-point and secondary end-point, respectively. Patients diagnosed as NPC were followed up primarily by telephone and periodic review in hospital. A total of 255 of 559 patients were followed up successfully. OS was defined as the period from the initial diagnosis to death regardless of NPC related or not or the last follow-up. The median follow-up time among the 255 patients was 33.5 months, ranging from 2.1 months to 151.2 months. DFS was defined as the period from the initial diagnosis to recurrence or metastasis. Follow-ups were ended in February 2019. Cox proportional hazards regression model. The logistic regression analysis was used to analyse the influencing factors of side effects after treatment. Receiver operating characteristic (ROC) curve was applied to determine optimal cut-off values and assess the predictive ability of prognostic indicators. 19 A P-value < .05 was considered statistically significant.

| Baseline characteristics of NPC and rhinitis patients
NPC and rhinitis were both common in men and younger patients (Table 1). Clinical parameters between NPC patients and rhinitis patients are shown in Figure 1. Most immune-inflammation indicators between two cohorts, such as PLR, NLR, SIRI and SII, were significantly different. To investigate the diagnostic significance of immunological indexes in NPC patients, ROC curve analysis was performed. As shown in Figure 2,

| The association between clinical indexes and haematological indicators in NPC patients
The association between haematological indicators and clinical characteristics in 559 NPC patients was shown in Table 2, and haematological indicators in a different circumstance, including therapy, TNM staging system and histopathological classification, were displayed in Figures 3-6. Significant differences in the haematological indicators were diverse in sex, age and metastasis status (Table 2).
Moreover, common differences in inflammation indicators (such as SII and PLR) in multiple comparative analysis can be observed . However, there were not significant differences in inflammation indicators in therapy and histopathological groups despite the difference in platelets in these groups (Figures 3 and 6).

| Influence of clinical indexes and haemograms on side effects
A total of 509 of 559 NPC patients received radiotherapy, but 2 patients of them were deficient in clinical data and therefore excluded in our study. Then, 507 patients were included in the study for side effects (Table S1).  (Table 4). And the independent risk factors for the skin pigmentation after radiotherapy included age, PLR, eosinophil and HCT ( Table 4). The independent risk factors for the dysphagia included eosinophil, HCT and PLR (Table 5), and the independent risk factors for the gastrointestinal reaction included sex, SIRI, M stage, eosinophil and HCT (Table 5). Haemoglobin, NLR and age were the independent risk factors for the innutrition (Table 6).
Age, eosinophil and HCT affected most side effects in the treatment of NPC patients, while T stage, N stage, histology, neutrophil and SII had no impact on these side effects.

| Clinical characteristics of immuneinflammation indicators in survival analysis
Finally, a total of 255 patients were enrolled in the study for survival

| Associations of immune-inflammation indicators with survival
The study took OS and DFS as the primary and secondary outcome,  (Table 9).

| D ISCUSS I ON
In the current study, we found that SIRI, SII, NLR, PLR, neutrophil, monocyte and RDW score were valuable for the prediction of both diagnosis and prognosis of NPC. And PLR was the best predictor of diagnosis of NPC.
Cancers can convert the peripheral matrix to promote progression. The changes involve recruitment of fibroblasts, migration of immune cells and formation of vascular networks. Tumour microenvironment (TME) comprises various cells and extracellular components. Excessive proliferation of cancer cells can stimulate the production of cytokines and chemokines, which attract immune cells to the TME and induce local immune inflammation. 21 Diem et al reported that NLR and PLR in the tumour microenvironment were associated with prognosis of lung cancer. 24 In addition, the circulating monocytes that play a major role in innate immunity may reflect the level of tumour-associated macrophages (TAMs), while TAMs can directly stimulate the growth, migration and metastasis of cancer cells. 25 Also, the platelet can promote tumour growth and metastasis owing to affecting cancer cells and other cells in the TME. 26 The different cell types in the TME communicate with each other to support cancer development; for example, SIRI and SII, the combination of NLR and monocyte and platelet, were associated with the prognosis of cancer patients. 19,27 Neutrophils can promote angiogenesis by pro-inflammatory cytokines, matrix metalloprotease 9 (MMP9) and VEGF, and can promote tumoral motility, migration and invasion. 28  Moreover, RDW is also a potential marker in tumour progression.
Mechanically, iron metabolism in red blood cells is affected by inflammatory factors, which induces the release of lots of immature red blood cells from the bone marrow in advance, and inflammatory factors also increase ineffective haematopoiesis in the bone marrow, which together induced a change in the RDW. 30 Wang et al reported that RDW and body mass index (COR-BMI) might serve as an inflammation-and nutrition-based indicator of prognosis in NPC. 31 Consistently, our results showed that RDW might help to predict the diagnosis and prognosis of NPC. The association between basophil and NPC has not been reported so far. In our study, NPC patients with high-score basophils had poor OS, which testified that basophil might participate in predicting the prognosis of NPC.
Besides, the NPC incidence of males is higher than that of females, and 50-to 60-year-olds are typical peaks. The ageing of the immune system may result in detrimental consequences on the response against cancers; then, the inflammatory status can promote immune suppression and cancer growth. 32 In our study, the incidence of NPC in males was three times higher than in females, and the incidence of patients who were under 60 years was three times higher than in those older than 60 years. And the risks of death of patients in the period of older than 60 years were 5.061 times higher than those in lower age.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
YP conceived and designed the manuscript. XZ, GL and YP acquired, analysed and interpreted the data and wrote and reviewed the manuscript. YL supervised the study.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.