Systemic Inammation with Sarcopenia Predict Survival in Patients with Gastric

Objective: The levels of platelet-related inammation indicators and sarcopenia have been reported to affect the survival of patients with cancer. To evaluate the prognostic inuence of platelet count (PLT), platelet–lymphocyte ratio (PLR), and systemic immune inammation index (SII), and SII combined with sarcopenia on the survival of patients with gastric cancer (GC). Methods: A total of 1131 patients with GC (811 men and 320 women, average age: 59.45 years) were evaluated. Receiver operating characteristic curves were used to determine the best cut-off values of PLT, PLR, and SII, and univariate and multivariate Cox risk regression models were used to evaluate whether SII is an independent predictor of overall survival (OS). The prognostic SS (SII-sarcopenia) was established based on SII and sarcopenia. Finally, a comprehensive analysis of the prognostic SS was performed. Results: SII had the strongest prognostic effect. The SII and OS of patients with GC were in an inverted U-shape (adjusted HR = 1.06; 95% CI: 0.95-1.18; adjusted P = 0.271). In patients with SII >1800, SII was negatively correlated with OS (adjusted HR = 0.57; 95% CI: 0.29-1.12; adjusted P = 0.102), however, there is no statistical difference. Interestingly, a high SS was associated with a poorer prognosis. The higher the SS score, the worse the OS (P<0.001). Conclusion: SII is an independent prognostic indicator of GC, and high SII is related to poor prognosis. A Higher SS score had worse survival. Thus, the prognostic SS is a reliable predictor of OS in patients with GC. Northern Europe(Sung et al., 2021). Most GC patients are diagnosed with advanced disease owing to lack of suitable biomarkers. Chronic inammation plays an important role in the occurrence of GC. The European Prospective Survey of Cancer and Nutrition study investigated the association between the inammatory potential of diet and the risk of GC in 476,160 subjects from 10 European countries. After 14 years of follow-up, the results showed that the inammatory potential of the diet is associated with an increased risk of GC(Agudo et al., 2018).


Introduction
Gastric cancer (GC) is the fth most common malignancy and the fourth leading cause of cancer-related mortality worldwide, with more than 1 million incident cases and 769,000 deaths recorded in 2020. The incidence of GC varies by region and is the highest in East Asia and Eastern Europe and lower in North America and Northern Europe (Sung et al., 2021). Most GC patients are diagnosed with advanced disease owing to lack of suitable biomarkers. Chronic in ammation plays an important role in the occurrence of GC. The European Prospective Survey of Cancer and Nutrition study investigated the association between the in ammatory potential of diet and the risk of GC in 476,160 subjects from 10 European countries. After 14 years of follow-up, the results showed that the in ammatory potential of the diet is associated with an increased risk of GC (Agudo et al., 2018).
Systemic in ammation plays a key role in the pathogenesis and progression of cancer. As such, the role of the tumor microenvironment in tumorigenesis has also attracted increasing attention. However, the interaction between tumors and in ammation is complicated. Preoperative hematological in ammation biomarkers including blood neutrophil, lymphocyte, monocyte, and platelet count (PLT); albumin levels; and their combinations have received increasing attention in recent years. Particularly, studies have reported that PLT plays an important role in in ammatory diseases. In addition, PLTs are involved in the occurrence and metastasis of cancer (Bambace & Holmes, 2011). Another study showed that the cancer type-speci c combination of PLT characteristics can be used to diagnose early-stage cancer (Sabrkhany et al., 2017). PLT-related in ammation biomarkers have also been identi ed to have a predictive value in various cancers (De Giorgi et  Sarcopenia is a syndrome that represents degeneration and systemic loss of skeletal muscle mass. According to recent surveys, the prevalence of sarcopenia is relatively high, ranging from 15% at the age of 65 to 50% at the age of 80. Patients with sarcopenia often have a higher incidence of infectious diseases, metabolic syndrome, insulin resistance, and cardiovascular disease. In patients with cancer cachexia, anorexia, malnutrition, and systemic in ammation, the catalytic effect of the metabolic state is enhanced, leading to sarcopenia. Therefore, sarcopenia is considered to be a manifestation of cancer cachexia (Fukushima, Takemura, Suzuki, & Koga, 2018). Recent studies have shown that sarcopenia has an impact on the prognosis of various cancers. The survival rate of patients with sarcopenia is signi cantly lower than that of patients with esophageal cancer (Jin et al., 2021), colorectal cancer (Xie et al., 2021), pancreatic cancer (Cho et al., 2021), lung cancer (Kawaguchi et al., 2021) without sarcopenia. In general, sarcopenia plays an important role in the prognosis of cancer patients.
Although platelet-related in ammation indicators or sarcopenia have a certain role in predicting the survival of cancer patients, the prediction effect needs to be improved. Hence, we explored the relationship between the combination of the two and the survival of patients with GC.

Study Design and Population
This observational cohort study analyzed the data of patients with GC who visited one of more than 40 clinical centers in China between 2012 and 2018. The exclusion criteria were missing or abnormal preoperative data (data=0) on neutrophil count, PLT, sarcopenia, tumor stage, and lymphocyte counts ( Figure 1). In total, 1131 GC patients were evaluated. This study was approved by the Institutional Review Board of each hospital (Registration number: ChiCTR1800020329) and was conducted in accordance with the Declaration of Helsinki. All participants signed a written informed consent.
Laboratory measurements included albumin levels, neutrophil counts, lymphocyte counts, and PLT counts. All blood tests were performed after fasting for at least 9 h before anti-tumor treatment within 48 h of the rst hospitalization. Body mass index (BMI) was calculated as BMI (kg/m 2 ) = weight (kg) / height^2 (m 2 ). PLR was calculated as PLR = PLT/lymphocyte counts, while SII was calculated as SII= neutrophil counts × PLT/lymphocyte counts. The SII-Sarcopenia Score (SS) is established based on SII and sarcopenia. Information on smoking status and alcohol and tea consumption was obtained using lifestyle questionnaires. All research results were reviewed and determined by an independent endpoint determination committee, whose members were blinded to the speci c tasks of the research team.
Pathological staging was according to the TNM staging system of the American Joint Committee on Cancer (8th edition)(Amin et al., 2017).

Statistical Analysis
The main endpoint was overall survival (OS), including death from any cause. Evidence of death was obtained from follow-up records. Predictive models were compared with the ROC curves and C-index. The chi-square test was used to compare the differences between the categorical variables in the baseline characteristics of the patients, and the t-test was used to compare continuous variables. The risk factors were modeled as continuous variables, and the optimal cut-off point was calculated using the Wilcoxon rank test. The chi-square test was used to model dichotomous and quartile SII and calculate the hazard ratios (HRs) and 95% con dence intervals (95% CIs) for the OS of patients with GC.
Adjusted variables included age, sex, tumor stage, drinking status, albumin level, neutrophil count, BMI, surgery, chemotherapy, and radiotherapy. The heterogeneity between subgroups was evaluated using Cox regression, and the interaction between SII and the subgroups was checked using the probability ratio. Kaplan-Meier curves of survival according to PLT, PLR, and SII levels were generated. All statistical analyses were performed using R software version 4.0.5 (Lucent Technologies).

Patient Characteristics
In total, 811 and 320 of the patients were male and female, respectively. The average age was 59.45 years. The optimal cut-off values predictive of survival were 270.50 for PLT, 149.57 for PLR, and 712.58 for SII, respectively (Supplemental Figure 1). Accordingly, low and high PLT, PLR, and SII were de ned as ≤270.50 and >270.50, ≤149.57 and >149.57, and ≤712.58 and >712.58, respectively. In total, 798 and 333 patients had low and high PLT, 640 and 491 patients had low PLR and high PLR, and 498 and 633 patients had low SII and high SII, respectively. The clinicopathological characteristics of the patients are shown in Table 1. The curves before and after adjustment showed an inverted U-shaped trend ( Figure 2).  Notes: Data are presented as hazard ratios (95% con dence intervals). The analyses are adjusted for age, sex, tumor stage, drinking status, albumin level, BMI, surgery, chemotherapy, and radiotherapy.
SII, systemic immune in ammation index; HR, hazard ratio; CI, con dence interval; P, probability; BMI, body mass index; Q, quarter To exclude the in uence of radiotherapy and chemotherapy on SII, we conducted a sensitivity analysis, and consistent results were obtained (Supplemental Table 1). In addition, we also performed COX regression analysis on patients undergoing surgery, radiotherapy and chemotherapy, and the results showed the same trend (Supplemental Table 2, Supplemental Table 3). High SII is related to poor prognosis. Using SII as a continuous variable for Cox regression analysis, low SII (≤1800) was signi cantly associated with poor prognosis (adjusted HR=1.16; 95% CI: 1.06-1.28; adjusted P=0.002). High SII (>1800) was also negatively correlated with poor prognosis, but there is no signi cant difference in this negative correlation (adjusted HR = 0.57; 95% CI: 0.29-1.12; adjusted P =0.102) (Supplemental Table 4).

Subgroup Analyses
The relationship between SII and OS among the different subgroups was evaluated using strati ed analysis (Figure 4) Figure 2).

Overall Survival
Kaplan-Meier curves of OS according to SII, sarcopenia, and SS are shown in Figure 5. High SII and sarcopenia were signi cantly related to the difference in OS in the univariate Cox proportional hazard regression (P<0.001). In addition, in the model comprising SII and sarcopenia showed that patients with high SII and sarcopenia had worse OS. Similarly, patients with a higher SS score also had worse OS (P<0.001).

Discussion
This multicenter cohort study found that among the PLT-related in ammation indicators PLT, PLR, and SII, SII has the best prognostic indication. Further analysis showed that patients with high SII is associated with poor OS in GC patients (Figure 3), and the HR gradually declined as SII increased to >1800. However, Cox regression analysis of SII showed that as the SII increased, survival worsened. A statistically signi cant positive correlation was found in the subgroup analysis with SII <1800 as the cut-off. Whereas in subgroup of SII >1800, HR=0.57, the statistically signi cant results were not shown. This might be due to the small number of patients with SII >1800. In addition, high SII combined with sarcopenia was associated with poor OS in patients with GC. The higher the SS score, the worse was the patient's OS ( Figure 5). These results have also been supported by a number of studies.
Cytokines are key components of the in ammatory process. PLTs play an indispensable role in the development and metastasis of cancer. Once activated, PLTs can bind to tumor cells through P-selectin, which can be found on the surface of PLTs and bind to the CD24 ligand. In addition, PLTs promote tumor growth and metastasis by releasing pro-angiogenesis and growth factors. Therefore, a large number of studies have focused on the prognostic implant of PLT-related in ammation indicators in cancer patients, but the results have been con icting. Therefore, other indicators that play an important role in predicting cancer survival have been evaluated ( . Notably, PLR and SII both include PLT PLR is considered to be a marker of endogenous residual anti-precancerous in ammation and procoagulant response in malignant tumors (Proctor et al., 2011). It is also considered to be a sensitive marker that could predict certain types of advanced cancer, treatment response, and prognosis (Zhou et al., 2014). In addition, an elevated SII indicates that a highly in ammatory tumor microenvironment. SII can be measured easily at an affordable cost using a reproducible method, making it a promising prognostic indicator in clinical applications (Dong et al., 2020).
Sarcopenia has also been con rmed to be related to poor prognosis in patients with cancer, thus making it a valuable prognostic indicator.
A meta-analysis in 2016 showed that most studies on SMI and prognosis of patients with cancer were published after 2012, and more than half of the studies were published after 2015. This indicates an increasing attention to the prognostic value of SMI in this population (Shachar, Williams, Muss, & Nishijima, 2016). Skeletal muscle loss after surgery has been con rmed to be signi cantly negatively correlated with adverse postoperative outcomes in patients with non-small cell lung cancer (Takamori et al., 2020). However, sarcopenia was measured using skeletal muscle area, which is different from our research. A retrospective analysis showed that SMI is an independent predictor of OS in patients with breast cancer (Hua et al., 2020). This association between SMI and cancer prognosis could be because certain tumors could induce systemic in ammation, and sarcopenia might be a re ection of more radical tumor metabolism. Cancer patients generally have varying degenerative diseases, causing loss of muscle mass, strength, and dysfunction. The occurrence of these degenerative diseases is in uenced by several factors including malnutrition, insu cient physical activity, comorbidities, and other factors directly related to pathophysiology and treatment-related toxicities.
There have been several studies on sarcopenia combined with in ammatory indicators. Chen  To our best knowledge, this is the rst cohort study to explore the relationship between SII combined with sarcopenia and OS in patients with GC. The advantage of this study is that we rst screened out the index SII that has the best predictive indication among platelet-related in ammation indexes, and combined with sarcopenia on this basis, we obtained a better predictive model SS. However, the following limitations should be also considered. First, the adopted anthropometric equations that had been validated in Chinese individuals to estimate muscle mass, rather than bioimpedance analysis (BIA) or DXA recommended by the European Working Group on Sarcopenia in Older People(Cruz-Jentoft et al., 2010) and Asian Working Group for Sarcopenia(L. K. Chen et al., 2014). However, DXA is expensive and the patients might have to expose to X-rays. There are very few BIAs in hospitals in mainland China, and the cut-off point for de ning low muscle mass on BIA among elderly Chinese individuals has not been established (Zeng et al., 2015). Second, only the PLT was recorded, and thus, PLT-related in ammation indexes did not included the mean PLT volume and PLT distribution width. Third, there was no record of diet and socioeconomic status, which might have affected muscle loss and cancer death. Fourth, this study only included the Chinese population and did not represent other ethnic groups. Future research needs to investigate whether reducing systemic in ammation and increasing muscle mass can prolong patient survival and the underlying mechanisms on their in uence in patients with GC.

Conclusion
Systemic in ammation and sarcopenia are relaed to the prognostic signi cance of GC patients. Among the PLT-related in ammation indicators PLT, PLR, and SII, SII has the best prognostic accuracy. In this study, SII independently predicted survival, thus making it a possible reliable indicator of GC prognosis. In addition, patients with both in ammation and sarcopenia have signi cantly worse survival than do patients without these impacts.   Relationship between SII and OS in patients with gastric cancer Notes: Use cox regression to analyze SII  The relationship between SII and the OS of patients with gastric cancer in different subgroups Notes: The cox regression model was used to calculate hazard ratios (HRs) and 95% con dence interval (CI). Each subgroup is adjusted for age, sex, tumor stage, drinking status, albumin level, BMI, surgery, chemotherapy, and radiotherapy. SII, systemic immune in ammation index; HR, hazard ratio; CI, con dence interval; P, probability; BMI, body mass index.