Prognostic value of systemic immune-inflammation index in patients with gastric cancer treated with immune checkpoint inhibitors

Abstract

BACKGROUND

Immune checkpoint inhibitors (ICIs) have improved the overall survival rates of advanced gastric cancer patients to some extent, but individuals respond differently; thus, an urgent need exists for precise predictive models. Systemic immune-inflammation index (SII) score, integrating peripheral blood neutrophils, T cells and platelets counts can comprehensively reflect changes in immunity and inflammation of human body. The prediction results of it are all quite accurate.

AIM

To research on the prediction capability of SII in patients with gastric cancer receiving immunotherapy.

METHODS

A total of 238 advanced gastric cancer patients who received ICIs have been retrospectively studied. Based on whether there was a deterioration in condition over six months (good prognostic group/poor prognostic group), as well as preoperative score II levels, patients were divided into low/high groups; cutoff: 593.995. Aneutropia, leucomasia and blood counts need to be tested weekly during this stage to assess the level of SII. Primary indicators are progression-free survival (PFS), etc.

RESULTS

The poor prognosis group was significantly higher in SII than the normal control group (673.42 ± 152.91 and 503.83 ± 138.74, respectively, P < 0.001). SII showed the best prediction accuracy in predicting progressive disease within 6 months [area under the curve (AUC) = 0.788]. The high SII group had a lower disease control rate (49.09% vs 69.53%, P = 0.001) and shorter PFS (7.83 months vs 8.52 months, P = 0.013). Based on multivariate analysis, a high SII was identified as an independent predictor of shortened PFS (hazard ratio: 1.896, P < 0.05). In receiver operating characteristic analysis, SII alone (AUC = 0.788) had a higher discriminative ability than other single clinical indicators; In addition, the highest prediction accuracy was obtained in combination mode (AUC = 0.856).

CONCLUSION

High SII scores have been confirmed to be necessary risk indicators, and by adding other clinical parameters, the diagnostic accuracy of progressive disease in ICIs-treated stomach cancer patients was improved significantly.

Key Words: Systemic immune-inflammation index; Gastric cancer; Immune checkpoint inhibitors; Prognosis; Biomarkers; Retrospective studies

Core Tip: The systemic immune-inflammation index (SII), calculated from routine blood counts, integrates neutrophils, lymphocytes, and platelets to reflect the host immune-inflammatory balance. This retrospective study of 238 advanced gastric cancer patients treated with immune checkpoint inhibitors demonstrates that a high baseline SII (≥ 593.995) independently predicts poor progression-free survival and lower disease control rates. Combining SII with clinical parameters significantly improves predictive accuracy for early progressive disease, suggesting SII as a practical and accessible biomarker for risk stratification in immune checkpoint inhibitors-treated gastric cancer.

INTRODUCTION

Gastric cancer has increasingly become a health problem for Chinese people over the years and its proportion of total cancers in recent years remains high. There has been some progress in the surgical method and post-operative therapy for advanced gastric cancer patients. The survival rate of these people remains relatively low at present[1-3]. Recently, there has been a surge in the emergence of immunotherapy that targets an immune checkpoint inhibitors (ICIs)-based new way for treating cancer by reactivating the host’s own immune system to eliminate aberrant cells[4]. Nevertheless, due to differences in sensitivity between patients using ICI drugs; predicting treatment effect and improving outcome by selecting key indicators during practical application is highly recommended.

In short, an immune-inflammatory microenvironment that triggers cancer-associated inflammation and increases the sensitivity of tumours to treatment. Currently recognised as the hallmark of malignancy that promotes tumour proliferation, angiogenesis, and metastasis through chronic inflammation inhibition in immune responses against cancer. Overall, in most cases, the inflammation-related biomarkers in stomach-type inflammatory cancers were correlated with risk factors; neutrophil-lymphocyte ratio; platelets/Leucocytes ratio lymphocyte-monocyte ratio. Indicators are reference values for evaluating the balance of pro-inflammatory and anti-inflammatory reactions in an organism’s immune defence against cancer to influence cancer biology and drug response. Often lacks recognizability and broad coverage. There is a shortage of unified indicators to demonstrate the complex relationship among systems-induced irritation and immune status more effectively[5-7].

ICIs reinvigorate tired T cells via the major programme of programmed cell death protein-1 (PD-1)/programmed cell death ligand 1 (PD-L1) blockade pathway, thereby restoring an active anti-tumour immunity. Nivolumab, pembrolizumab among the PD-1 drugs have shown varying degrees of effectiveness against gastric cancer in large numbers of phase III clinical trials. Checkmate 649, the first clinical trial that showed some benefit over combined administration of nivolumab plus chemotherapy vs monotherapies for all patients receiving chemotherapy or previous studies, has been described as one such landmark finding in recent years. KEYNOTE-811 study found that the combination therapy with pembrolizumab added to trastuzumab and chemotherapy demonstrated a higher objective response rate in human epidermal growth factor receptor 2 (HER-2)-expressing gastric cancer patients compared to the monotherapy[8]. The combination of neoadjuvant immunotherapy and chemotherapy showed a higher pathologic complete response rate than either treatment alone; It was still advanced-stage for localised early gastric cancer[9,10]. Nevertheless, a meta-analysis incorporating 17 randomized controlled trials suggested that PD-1/PD-L1 inhibitors did not substantially boost patient survival results in subsequent-line therapies for gastric cancer, indicating a temporal-based benefit of immunotherapy[11]. As shown in the above example, ICI effects are complicated; some malignant tissues have high PD-L1 expression and evade immunosurveillance. In addition, various immunosuppressive activities occur in the tumour microenvironment. For instance, a high infiltration degree of regulatory T-cells and myeloid-derived suppressor cells (MDSCs). At the same time, there has been an increase in the expression level of alternative immune check point pathways, such as galectin-9/tim-3)[10]. Therefore, under this form of reducing the uncertainty when selecting treatment options for cancer patients with particular biological features, it has already begun.

Systemic-immune-inflammatory parameters have been included in recent studies as additional or complementary indicators to be used when predicting disease outcomes besides the traditional ones. The difference between neutrophil lymphocyte ratio and platelet-to-lymphocyte/Lymphocyte-monocyte) ratio; systemic immune-inflammation index (SII) integrates various immunological-and-inflammation-based markers to offer a relatively reliable assessment on whole-body’s inflammatory condition[12,13]. Some previous studies have also shown that the influence of SII on colorectal cancer and hepatocellular carcinoma may be transferable to other malignant tumours[14,15]. Gastric cancer has some degree of association with the level of sensitivity index, but it is not a complete picture in general. SII can reflect both immune activation and inhibition, making it suitable for use in combination with immunotherapies that aim to achieve a good balance between these two types of responses to have more significant anti-tumour effects[16,17].

Although there are many applications for SII. Nevertheless, its effectiveness must be confirmed through clinical observation in combination with those who have received ICIs. Some studies have shown that using GII for predicting the survival of gastric cancer patients. However, few are related to therapy outcomes after surgery or chemotherapy alone. There is insufficiently structured clinical data to assess whether SII independently has predictive value in determining the therapeutic effect of patients with gastric cancer who have undergone immunotherapy by combining other factors, such as PD-L1 expression and tumor burden, alone. Due to difficulties in determining how the mechanism of immunomodulation and the individual differences in gastric cancer affect their responsiveness to ICIs as personalized treatment references. Also identify the high-risk population of this group who responded well to ICIs according to their SII score values in future treatments to improve patient care and treatment results. Thus, based on existing clinical data, exploring the association between initial SII scores and treatment outcomes [including disease control rates (DCRs) and progression-free survival (PFS)] in advanced gastric cancer patients receiving ICI therapy, also hope to find that SII might be considered as an independent prognostic factor. Thus, providing new reference for clinicians to identify high-risk individuals suitable for immunotherapies. Therefore, when assessing SII beyond being a tool for prediction purposes; rather, as an assistant that may guide treatment decisions.

MATERIALS AND METHODS.

Study materials

A retrospective study of 238 patients over the age of seventy-five with advanced gastric cancer who received ICIs-based treatment between January 1, 2022 and June 1, 2024 at Henan Provincial People’s Hospital. The research has been approved for ethics by the Ethics Committee of Henan Provincial People’s Hospital, and individual informed consent is not required.

Inclusion criteria: (1) Between 18 years old and 75 years old; (2) Definite pathological diagnosis of gastric cancer confirmed by histopathology examination[18]; (3) Clinically staged as stage III or higher according to the Eighth Edition of the American Joint Committee on Cancer’s Gastric Cancer Staging System (Eighteen)[19]; (4) Eastern Cooperative Oncology Group (ECOG) performance status score ≤ 2; (5) Has received more than two courses of immunotherapy, which includes PD-1/PD-L1 inhibitors, guided by a valid treatment plan (monotherapy or combination). Integrated therapy combines chemotherapy, including XELOX, FOLFIRX, SOXA/LOCA+platinum compound, etc.[20], anti-HER-2 targeted treatment for HER-2-positive patients[21,22], combination of angiogenesis inhibitor, such as apatinib or others target[23], dual Immunotherapy synergy[20], paired with other targeted drugs (CLDN18.2-targeting drug combinations)[21]; (6) Perform peripheral blood routine test within one week before taking a cycle of administration to get the lab results include count neutrophils, lymphocyte, platelet which could be applied for calculation SII index; (7) Able to cooperate take part at least six consecutive months follow-up treatment; and (8) Have all medical records no missed any entry items.

Inclusion and exclusion criteria: (1) Coexistent of primary malignant tumours; and (2) Active infection requiring systemic antibiotics, antifungals or anti-virals during the course of immunotherapy, or a long-term state resulting in impaired peripheral-blood cell counts evaluation by the investigator (e.g., due to haemopoietic suppressive syndromes or hematological malignancies), which affects the accuracy of immune-checkpoint inhibitor efficacy assessment. No serious and unmanageable heart damage has been confirmed at a high level of medical evidence provided in this study (such as New York Heart Association class III-IV cardiac insufficiency, unstable angina pectoris, etc.) or severe muscle disease necessitating medication.

Definitions

Peripheral neutrophils, lymphocytes and thrombocytes were counted using a high-quality haematology analyser (SYSMEX XN-9100); System MEX Co., Ltd. (Japan). Platelet count × neutrophil count/Lymphocyte ratio, > 1.5 as weakly impaired; < 1.0 as severely impeded or infected recently[24].

Combined positive score (CPS) ≥ 1 in the upregulated PD-L1 group at our institution. The CPS was determined using the following method: (1) The total percentage of positively viable neoplastic cells (including partial or complete membranous staining); and (2) Plus positively immunoreactive lymphocytes and mononuclear phagocytes (including membrane or cytoplasmic staining), which were scored on a scale ranging from 0 points to 100 points (if the calculated result exceeded 100, it should be recorded as 100)[25].

Based on RECIST 1.1 criteria, determine whether patients have achieved an objective response based on the diagnostic opinions of more than five years’ experience from a qualified cancer-image diagnosis personnel in this field. The treatment result is rated as a complete cure, partial response (PR), and stable condition [stable disease (SD)], progressive disease (PD). The DCR = (cured completely + mildly cured + unchanged)/total number of patients × 100%[26].

As a key indicator for predicting disease-free survival, the time point from receiving immune therapy until the onset of disease according to RECIST 1.1 criteria; or death caused by PD, such as progression, etc.; it belongs to one of the stages of this trial’s observation period. The ones whose PD or death occurred after the end of observation; That is, the follow-up-enddate-treatment-starteddate was taken as a censored event during the follow-up period and included in the analysis.

Follow-up

At least six months following the administration of immunotherapy for all patients who underwent gastrectomy due to gastric cancer were contacted via outpatient visit or telephone inquiry. The end of the follow-up period was January 1, 2025; the median follow-up time was 11.4 months (range: 6.0-24.8 months). The follow-up content included the following contents: (1) Survival status (lived/dead); (2) Whether there has been a change in the course of disease; (3) The grade of adverse event based on CTCAE-5.0 standard, etc.; and (4) Information about subsequent anti-tumour therapy after occurrence of progression or other adverse events were recorded. After ICI therapy, outpatient follow-up examinations were scheduled every two cycles (at the time of efficacy evaluation) to include a comprehensive physical examination, lab tests such as complete blood counts, liver and kidney functions, cancer-related biomarkers, etc., and computed tomography/magnetic resonance imaging scans. Within three months of completing the treatment (because of PD, adverse events or voluntarily discontinued) carried out by a professional physician, telephone follow-up every month is needed, including patient symptoms, ECOG grade, new problems have appeared, and recent medical history. From three months to six months old, telephonic check-ups occurred biweekly during this period. If patients are still living and do not have PD after six months, a phone visit will be arranged every three months until the last follow-up ends. Lost-to-follow-up patients were supplemented with their survival status from the health insurance system, death registration system or hospitalisation records. The data for patients without events until the last examination was censored on the day after they contacted us.

Research group

On this basis, in combination with whether progression has occurred within six months after initiating immune checkpoint inhibitor therapy. There are 238 patients who have not progressed in six months as part of the good-prognosis group (no PD during this period). Among them, there were 143 cases that met these conditions. Based on the determination of the SII value in the previous 1 week prior to ICI administration using receiver operating characteristic (ROC) curves. Then divide them into two groups according to whether their SII values exceeded this optimal threshold (best threshold: 593.995). Therefore, there would be 128 people classified as a low-SII group and 110 persons grouped as high-SII group respectively.

Statistical analysis

Using SPSS version 29.0 by SPSS Inc. to perform statistical analyses in this study. All of the continuous variables met the assumption of normality according to the Shapiro-Wilk test; therefore, they will be expressed as mean ± SD. Group comparisons of the continuous variable used an independent samples t-test. Categorical data were presented in the form of n (%) to compare differences among Groups through χ² tests. Multivariate logistic regression analysis was performed to determine which of the following were not correlated with having PD after 6 months in patients receiving ICI therapy. Next, ROC analysis was performed to find the best threshold for SII that maximised the Youden Index. Kaplan-Meier curves were plotted for the patient survival, and then using multivariate Cox proportional hazard regression, we investigated the risk factors that affected PFS among patients who received treatment with ICI. P < 0.05 met the criteria for statistical significance.

RESULTS

Comparison of the poor and good-prognosis group

Baseline characteristics: Compared with patients in the good-prognosis group, those in the poor-prognosis group had higher scores on the ECOG scale (χ² = 12.643; P = 0.002), and a significantly greater proportion scored 2 than that of the excellent category (26.32% vs 9.09%, Table 1). Moreover, clinical staging differed significantly (χ² = 10.659, P = 0.001), with stage IV being more common in the poor prognosis group (83.16% vs 63.64%). There were more distant metastases in the poor-prognosis group (χ² = 8.636; P < 0.05), and 62.11% of them exceeded two sites, while only 42.66% of those in the best outcome group had this many. The treatment plan differed significantly (χ² = 8.325; P = 0.004). A higher percentage of patients in the favourable prognosis group received combined therapy (58.04% vs 38.95%). PD-L1 expression (CPS) had a significant difference (χ² = 8.692, P = 0.003). The proportion of patients with positive cells (≥ 1) in the poorly prognostic group was 74.74%, while that in the well-prognosis group was only 55.94%. Gender, age, body mass index and history of gastric cancer resection operation were all the same in both groups (P > 0.05).

Table 1 Comparison of baseline characteristics of gastric cancer patients in the poor prognosis and good prognosis groups, n (%)/mean ± SD.

Parameters	Good prognosis group (n = 143)	Poor prognosis group (n = 95)	t/χ²	P value
Gender			0.163	0.687
Female	52 (36.36)	37 (38.95)
Male	91 (63.64)	58 (61.05)
Age (years)	62.34 ± 4.15	63.28 ± 3.87	1.756	0.080
Body mass index (kg/m2)	22.15 ± 3.02	21.89 ± 2.24	0.756	0.450
Eastern Cooperative Oncology Group			12.643	0.002
0	72 (50.35)	38 (40.00)
1	58 (40.56)	32 (33.68)
2	13 (9.09)	25 (26.32)
Clinical staging			10.659	0.001
IV	91 (63.64)	79 (83.16)
III	52 (36.36)	16 (16.84)
Number of distant metastases			8.636	0.003
≥ 2	61 (42.66)	59 (62.11)
< 2	82 (57.34)	36 (37.89)
Previous gastric cancer resection surgery			0.692	0.406
Yes	65 (45.45)	38 (40.00)
No	78 (54.55)	57 (60.00)
Treatment regimen			8.325	0.004
Combination therapy	83 (58.04)	37 (38.95)
Monotherapy	60 (41.96)	58 (61.05)
Programmed cell death ligand 1 overexpression (combined positive score)			8.692	0.003
≥ 1	80 (55.94)	71 (74.74)
< 1	63 (44.06)	24 (25.26)

Laboratory items: In terms of laboratory results among patients with gastric cancer who have a good or bad outcome, the platelet count in those classified as having a poor outcome was found to be significantly higher compared to those without an adverse event (230.71 × 10⁹/L ± 74.25 × 10⁹/L for “poor” condition; 207.46 × 10⁹/L ± 68.33 × 10⁹/L for “good” condition, t = 2.482, P = 0.014; Table 2). Neutrophil count was found to be higher in the poor prognosis group (4.32 × 10⁹/L ± 1.37 × 10⁹/L) than in the good prognosis group (3.91 × 10⁹/L ± 1.12 × 10⁹/L, t = 2.442, P = 0.016). The lymphocyte count was significantly higher in the good prognostic group (1.61 × 10⁹/L ± 0.38 × 10⁹/L), while that of the poor prognostic group was relatively low at 1.48 × 10⁹/L ± 0.41 × 10⁹/L (t = 2.398, P = 0.017). SII showed a significant difference between groups; it was much larger in the poor prognosis group (673.42 ± 152.91), and significantly higher compared to that of the other two subgroups (503.83 ± 138.74; t = 8.864, P < 0.001).

Table 2 Comparison of laboratory indicators of gastric cancer patients in the poor prognosis and good prognosis groups, mean ± SD.

Parameters	Good prognosis group (n = 143)	Poor prognosis group (n = 95)	t value	P value
Platelet count (× 109/L)	207.46 ± 68.33	230.71 ± 74.25	2.482	0.014
Neutrophil count (× 109/L)	3.91 ± 1.12	4.32 ± 1.37	2.442	0.016
Lymphocyte count (× 109/L)	1.61 ± 0.38	1.48 ± 0.41	2.398	0.017
Systemic immune-inflammation index	503.83 ± 138.74	673.42 ± 152.91	8.864	< 0.001

Multivariate logistic regression analysis: Several important factors were found to be related to PD occurring within six months after ICIs use in stomach cancer patients through multivariate logistic regression analysis (Table 3). Higher ECOG grade, clinical stage and the number of distant metastasis were found to be risk factors. The odds ratios (ORs) and corresponding 95%CI were respectively as follows: (1) ECOG status (OR = 1.740, 95%CI: 1.094-2.770, P = 0.019); (2) Clinical Staging (OR = 2.459, 95%CI: 1.122-5.391, P = 0.025); and (3) Distant metastatic sites (OR = 1.982, 95%CI: 1.014-3.875, P = 0.046). Another risk factor for the PD-L1 expression score ≥ 1 was (OR = 2.234; 95%CI: 1.089-4.586, P = 0.028). By contrast, combination therapy was identified as a protective factor against PD (OR = 0.486; 95%CI: 0.249-0.951; P = 0.035). SII had a high degree of correlation with an elevated risk of PD (OR = 1.006, 95%CI: 1.004-1.009, P < 0.001).

Table 3 Multivariate logistic regression analysis of the risk factors for progressive disease within six months in gastric cancer patients receiving immune checkpoint inhibitors.

Parameters	Coefficient	SE	Wald Stat	P value	OR	OR 95%CI lower	OR 95%CI upper
Eastern Cooperative Oncology Group (2)	0.554	0.237	2.337	0.019	1.740	1.094	2.770
Clinical staging (IV)	0.900	0.400	2.247	0.025	2.459	1.122	5.391
Number of distant metastases (≥ 2)	0.684	0.342	1.999	0.046	1.982	1.014	3.875
Treatment regimen (combination therapy)	-0.721	0.342	-2.108	0.035	0.486	0.249	0.951
Programmed cell death ligand 1 overexpression (combined positive score ≥ 1)	0.804	0.367	2.192	0.028	2.234	1.089	4.586
Systemic immune-inflammation index	0.006	0.001	5.566	< 0.001	1.006	1.004	1.009

OR: Odds ratio.

ROC analysis: In the ROC analysis of risk factors for PD within 6 months in patients with gastric cancer who received ICIs, SII had the highest discriminatory ability, with an area under the curve (AUC) of 0.788; its sensitivity was 0.705, specificity was 0.727, and Youden’s index was 0.432 (Table 4). ECOG (2) had an AUC of 0.590, with sensitivity and specificity of 0.263 and 0.909, respectively. The clinical stage IV (AUC = 0.598), with sensitivities of 0.832 and specificities of 0.364, respectively. The Number of distant metastases (≥ 2) has an AUC of 0.597, its sensitivity is 0.621 and specificity is 0.573. Among these combinations, the one whose AUC is 0.595 has a sensitivity of 0.611 and a specificity of 0.58. The AUC for PD-L1 expression (CPS ≥ 1) was 0.594; the sensitivity was 0.747, and the specificity is 0.441. The combined ROC analysis further improved the predictive accuracy, yielding an AUC of 0.856 (also not shown separately) (Figure 1).

Figure 1 Combined receiver operating characteristic curve for predicting progressive disease within six months in gastric cancer patients receiving immune checkpoint inhibitors. AUC: Area under the curve; ROC: Receiver operating characteristic.

Table 4 Receiver operating characteristic analysis of the risk factors for progressive disease within six months in gastric cancer patients receiving immune checkpoint inhibitors.

Parameters	Best threshold	Sensitivities	Specificities	Area under the curve	Youden index	F1 score
Eastern Cooperative Oncology Group (2)	0.500	0.263	0.909	0.590	0.172	0.376
Clinical staging (IV)	0.500	0.832	0.364	0.598	0.196	0.596
Number of distant metastases (≥ 2)	0.500	0.621	0.573	0.597	0.194	0.549
Treatment regimen (combination therapy)	0.500	0.611	0.580	0.595	0.191	0.344
Programmed cell death ligand 1 overexpression (combined positive score ≥ 1)	0.500	0.747	0.441	0.594	0.188	0.577
Systemic immune-inflammation index	593.995	0.705	0.727	0.788	0.432	0.667

Comparison between high and low SII groups

Basel information: Table 5 shows that there is a notable difference in some parameter values between the gastric cancer patients’ group in the low-SII and high-SII clusters, respectively. The ECOG performance index distribution was not uniformly distributed (χ² = 8.111; P = 0.017), where stage IV was more prevalent in the high SII group (80.00%) than in the low SII group (64.06%). Clinical staging also showed a notable difference: χ² = 7.363, P = 0.007; stage IV of the tumour was comparatively more likely to occur in patients whose scores were higher than SII (80.00%) than those lower than it (64.06%). The number of distant metastasis was also an important difference between groups (χ² = 6.151, P = 0.013), with patients in the high SII group having a greater incidence of having two or more metastases (59.09%) than those in the low SII group (42.97%). Gender, age, body mass index, history of gastric cancer resection surgery, treatment regimens and the level of PD-L1 overexpression were not significantly different from other groups (P > 0.05).

Table 5 Comparison of baseline information of gastric cancer patients in the high and low systemic immune-inflammation index groups, n (%)/mean ± SD.

Parameters	Low SII group (n = 128)	High SII group (n = 110)	t/χ²	P value
Gender			0.054	0.816
Female	47 (36.72)	42 (38.18)
Male	81 (63.28)	68 (61.82)
Age (years)	62.45 ± 3.97	63.02 ± 4.05	1.101	0.272
Body mass index (kg/m2)	22.18 ± 3.01	21.87 ± 3.08	0.790	0.430
Eastern Cooperative Oncology Group			8.111	0.017
0	67 (52.34)	43 (39.09)
1	48 (37.50)	42 (38.18)
2	13 (10.16)	25 (22.73)
Clinical staging			7.363	0.007
IV	82 (64.06)	88 (80.00)
III	46 (35.94)	22 (20.00)
Number of distant metastases			6.151	0.013
≥ 2	55 (42.97)	65 (59.09)
< 2	73 (57.03)	45 (40.91)
Previous gastric cancer resection surgery			1.460	0.227
Yes	60 (46.88)	43 (39.09)
No	68 (53.12)	67 (60.91)
Treatment regimen			0.014	0.904
Combination therapy	65 (50.78)	55 (50.00)
Monotherapy	63 (49.22)	55 (50.00)
Programmed cell death ligand 1 overexpression (combined positive score)			0.045	0.831
≥ 1	82 (64.06)	69 (62.73)
< 1	46 (35.94)	41 (37.27)

SII: Systemic immune-inflammation index.

Efficacy assessment: Compared to the effect of therapy for patients with gastric cancer in both the low and high SII groups; the proportion including PR or SD was statistically significant (χ² = 10.306, P = 0.001; Table 6). Among those who reached the DCR criteria in both groups: (1) Low SII group 69.53%; and (2) High SII group 49.09%. In the low SII group, PR cases accounted for 12.50% and SD cases for 57.03%, while in the high SII group, the PR rate was 6.36% and the SD rate was 42.73%. The proportion of patients with PD in the high SII group (50.91%) was significantly higher than that in the low SII Group (30.47%).

Table 6 Comparison of therapeutic response of gastric cancer patients in the high and low systemic immune-inflammation index groups, n (%).

Parameters	Low SII group (n = 128)	High SII group (n = 110)	χ²	P value
Disease control rate (PR + SD)	89 (69.53)	54 (49.09)	10.306	0.001
PR	16 (12.50)	7 (6.36)
SD	73 (57.03)	47 (42.73)
Progressive disease	39 (30.47)	56 (50.91)

PR: Partial response; SD: Stable disease; SII: Systemic immune-inflammation index.

Prognostic assessment: A significant difference was observed in the comparison of PFS between patients with gastric cancer in the low and high SII groups (t = 2.492, P = 0.013; Figure 2). The median PFS of patients in the low SII group was 8.52 months; compared to those in the high SII group, it reached up to 7.83 months and had increased by 0.69 months in median PFS.

Figure 2 Kaplan-Meier plots of progression-free survival of gastric cancer patients in the high and low systemic immune-inflammation index groups. A: Survival curve; B: Number of patients. SII: Systemic immune-inflammation index.

Multivariate Cox proportional hazards model analysis: The following several risk factors influencing PFS in gastric cancer patients receiving ICI therapy were found to have a statistically significant association after controlling for other variables (Table 7). ECOG [P = 0.018; hazard ratio (HR) = 1.642; 95%CI: 1.088-2.478]; clinical stage (IV: P = 0.001, HR = 2.149; 95%CI: 1.361-3.394), and the number of distant metastases ≥ 2 (P = 0.032; HR = 1.593; 95%CI: 1.042-2.435) showed poorer PFS. As well as that, a high SII (≥ 593.995) was also a notable risk factor (P = 0.005; HR = 1.896, 95%CI: 1.214-2.961). On the other hand, treatment regimen (combination therapy) and PD-L1 overexpression (CPS ≥ 1) did not show significant associations with PFS (P = 0.076 and P = 0.124, respectively; all P > 0.05).

Table 7 Multivariate Cox proportional hazards regression analysis of risk factors affecting progression-free survival in gastric cancer patients treated with immune checkpoint inhibitors.

Parameters	P value	Hazard ratio	95%CI
Eastern Cooperative Oncology Group (2)	0.018	1.642	1.088-2.478
Clinical staging (IV)	0.001	2.149	1.361-3.394
Number of distant metastases (≥ 2)	0.032	1.593	1.042-2.435
Treatment regimen (combination therapy)	0.076	0.712	0.491-1.032
Programmed cell death ligand 1 overexpression (combined positive score ≥ 1)	0.124	0.803	0.608-1.061
High systemic immune-inflammation index (≥ 593.995)	0.005	1.896	1.214-2.961

DISCUSSION

Among them, the baseline characteristics of ECOG grade, clinical staging and the number of distant metastasis were significantly different between the good-prognosis group and the poor-prognosis group at their earliest stage. Patients with higher ECOG scores generally have more developed diseases, other illnesses, etc., and thus are less able to tolerate intensive therapies[27,28]. There are many remote metastases present; therefore, there is a higher risk of poor treatment outcomes for this patient. The tumour volume is positively correlated with the systemic inflammatory response score caused by elevated cytokines; therefore, their prognosis might be poor. Systemic inflammation will further weaken the body’s immune function, and thus trigger or enhance a chain reaction of immunodeficiency[29,30].

Another aspect of clinical grading also differed between the two groups. Advanced stages generally refer to a greater degree of tumour dissemination and higher systemic inflammation due to tumours reaching an advanced stage. Therefore, these tumours release more inflammatory substances that weaken the body’s own defences. At this time, the weakening of immune functions may promote the occurrence and development of neoangiogenesis to provide nutrition support for tumor growth[31,32].

Provided additional support for the system-wide inflammatory response impacting patients’ outcomes. The number of platelets was higher in the bad prognosis group, which may have been induced by pro-inflammatory cytokine releases following inflammatory reactions. Neutrophils are significantly elevated (indicating ongoing inflammation); lymphocytes have decreased levels due to a compromised immune response. SII, which combines all of them, was significantly greater in the poor-prognosis group. This conclusion also asks whether this is merely a reflection of the degree of improvement in prognosis or a cause that drives changes in prognosis outcomes? Based on previous research, persistent chronic inflammatory state that does not occur necessarily in conjunction with tumour growth has become a trigger factor for the immunosuppression of cancer and drug resistance by different means[33,34].

SII increases directly as an indicator of a worse prognosis in tumour-immune microenvironment remodelling. With a higher SII value, there will be more increases in neutrophils and platelets as well as reductions in lymphocytes; thus, resulting in the formation of an immunosuppressive microenvironment. Neutrophils can transform into MDSCs that reduce T-cell activation and proliferation through the release of arginase-1 and inducible nitric oxide synthase. Meanwhile, activated platelets release transforming growth factors-beta and vascular endothelial growth factors to exacerbate immunosuppression and promote angiogenesis[35]. Platelets can shield circulating tumour cells through the formation of “platelet-tumour-cell aggregates”, making them less susceptible to immune detection and physical destruction, thus promoting metastasis[36]. In this study, a larger proportion of patients with multiple metastases (59.09% vs 42.97%) and a shorten PFS in the high-SII group were found as clinical outcomes based on these pathways.

A significant increase in SII is associated with active key proinflammatory signalling pathways. Systemic chronic inflammatory response often accompanies an enhanced activation state of related signaling pathways, including nuclear factor-kappa B (NF-κB) and signal transducer and activator of transcription 3 (STAT3) among others, that are key factors affecting tumourigenesis and immune cell behaviour. The activation of the NF-κB pathway increases PD-L1 expression in tumour cells and makes these cells less immune-difficult. STAT3 pathway promotes the growth and survival of tumour cells, as well as angiogenesis; it inhibits dendritic cell differentiation and activation functions and reduces adaptive immunity response[29,30]. The activation of these molecular mechanisms explains why patients in the high SII group not only had a lower DCR (49.09% vs 69.53%) but were also more prone to early PD. It should be noted that recently, a study involving patients with gastric cancer who received nivolumab indicated that the combination of SII and the revised Glasgow Prognosis Score was able to more accurately forecast early-treatment failure; specifically, those in higher-risk categories had an increased probability[37].

Multivariate logistic regression analysis revealed that the following risk factors were associated with progression within six months: (1) Higher performance status (ECOG) score; (2) Advanced clinical stage; and (3) More distant metastasis sites. Factors PD-L1 overexpression and therapy type are important. Tumour PD-L1 expression inhibits the activity of T-cells to induce tumour escape via this mechanism. The reasons are: Higher PD-L1 expression is associated with poor prognosis[38]. On the other hand, combined use demonstrated protection, possibly due to attacking multiple avenues of tumour proliferation and immune suppression at once. All parameters had better performance in ROC analysis except for SII. This result indicates that SII can reflect all dimensions of the immunological and inflammatory status of patients at present[39].

The comparisons of the low SII group and high SII group showed significant changes at the beginning of treatment. A larger positive correlation coefficient value (SII scores), indicating more diffuse lesion spread, thus increased the correlation of SII with the severity grade. Chronic inflammation triggered by high levels of SII can activate the signal pathway via NF-κB and STAT3, promoting tumour growth and drug-resistance development. The above-mentioned pathways enhance the survival, proliferation and angiogenesis of cancer cells through multiple mechanisms, thereby enhancing tumour invasiveness[40,41].

Furthermore, the high SII group had lower DCRs and shorter PFS. Referring to the obtained data shows that SII is a factor reflecting both its present condition and development direction of exacerbation in inflammation. Chronic inflammation persists in activating immune-suppressing cells and cytokines for extended periods, thus promoting tumour development and limiting the efficacy of immunotherapy[42]. Notably, an increased SII is a “consequence” of an immunosuppressive state and a “cause” that further exacerbates immune dysfunction. Under chronic inflammation, there is a bias in bone marrow for the differentiation of immature neutrophils and MDSCs. These cells after entering the bloodstream and tumour site will again inhibit immune functions by multiple pathways to form an autonomous vicious circle[43]. The above understanding has important clinical application value: For patients with high SII, ICIs alone are unlikely to fully cross the immune barrier. In other words, combine them with anti-inflammatory therapies and intervention strategies that target MDSCs to reconstruct the tumour micro-environment favourable for immunotherapies.

Multivariate Cox proportional hazards regression analysis revealed that several factors were correlated with PFS. Higher ECOG grade, more advanced clinical stage, and larger extent of distant metastasis correlated with shorter PFS. Advanced clinical presentation is usually a large tumor associated with extensive metastasis and thus causes severe systemic inflammation. The inflammatory response activates multiple Signal Transduction Pathways to promote tumour growth and suppress immunity, thus shortening lifespan. In addition, the high-SII group was found to be at higher risk of decreased PFS (HR = 1.896, P < 0.05). A high value of SII might reflect the presence of more immunosuppressive cells, including MDSCs and regulatory T-cells that weaken anti-tumour immunity. Combinedly, this setting does not support an efficient immune response and thus leads to a shortened PFS[33,44]. This HR is consistent with the results of recent studies showing that immunotherapy for gastric cancer has certain effects: According to a study on peroperative FLTLO chemotherapy, it was found that the high-SII group had a death risk-HR of 1.88 (95%CI: 1.36-2.89)[33]; in another study involving patients with metastatic gastric cancer, SII, together with the PD-L1 CPS score, was confirmed as an independent predictor of efficacy for nivolumab combined with chemotherapy[34]. Consistent with the above evidence, it can be predicted that SII is applicable for predicting prognosis in gastric cancer patients receiving immunotherapy.

ROC analysis showed that the sensitivity of SII screening is in a mid-high range. Although the parameter scores of ECOG and clinical stage only partially predicted that SII showed stronger discriminating power than either one of them. The above-mentioned results show that, in summary, these features of SII are all associated with systemic inflammation and immune dysfunction. Given its higher sensitivities and specificities, the SII is suitable for identifying individuals with a high likelihood of responding positively to immune checkpoint inhibitor therapies, enabling personalised therapeutic approaches. In addition, when used in conjunction with other biomarkers, SII is expected to enhance the prediction effect and precision of patients’ classification[45]. Based on these studies and previous research findings, the value of SII is generally accepted to have a strong predictive function in clinical application for patient prognosis after lung cancer surgery. It has been replaced by more functions as both indicators for quantifying inflammatory responses and measures to assess biologic events driving tumour proliferation and treatment resistance. Thus, in terms of creating individualisation plans according to SII data; that is what this paper contributes. Given that the patient has a higher SII score, additional rigorous combinations and adjuncts with anti-inflammatory effects should be considered during treatment.

However, there are deficiencies in the current study; it is an old-style study with the possibility of selection bias. There is still residual confounding after controlling for all the variables. One of the key aspects that this study has not systematically gathered and analysed in terms of patients’ co-administered medication, including antibiotics, proton pump inhibitors and corticosteroids. Consequently, the failure to incorporate these potential confounding factors into the multivariate analysis may introduce bias into the interpretation of the study results. In the future, further studies are needed to replicate these results in large-scale, long-term prospective cohort designs; also, more detailed biological pathways connecting SII with ICI response need exploration. Introducing other new biomarkers into the SII system to increase predictive power and guide individualised therapy.

This study highlights the importance of SII as a prognostic indicator in patients with gastric cancer receiving ICIs. To clarify the regulatory mechanism of systemic inflammation-tumorigenesis-metastasis can open up a path to develop novel target anti-inflammation therapy for cancers. Still lacking research efforts that focus on enhancing hospitalisation application efficiency and thereby positively impacting patients’ recovery rates.

CONCLUSION

The investigation into the prognostic value of SII in patients with gastric cancer treated with ICIs suggests that SII may serve as a potential biomarker for predicting patient outcomes. Higher SII levels appear to be associated with more advanced disease stages and greater tumor burden. SII demonstrated superior discriminatory power over other clinical parameters in identifying patients at increased risk of PD. SII was identified as an independent risk factor of PFS, further emphasizing its significance in assessing patient prognosis. The study underscores the importance of SII in reflecting systemic inflammation and immune function, potentially guiding more precise therapeutic decisions.