Spatial immune cell architecture after gastric cancer surgery predicts prognosis

Abstract

BACKGROUND

Gastric cancer remains a major global health burden with high mortality due to late-stage diagnosis and limited treatment efficacy. Emerging evidence demonstrates that the tumor microenvironment, particularly immune cell infiltration patterns and hypoxia-induced immunosuppression, plays a critical role in cancer progression. This study investigates the spatial immune cell architecture in postoperative gastric cancer tissues to identify prognostic biomarkers and guide personalized immunotherapy strategies.

AIM

To investigate the spatial distribution pattern, density, and proportional features of important immune cells in the tumor microenvironment following radical gastrectomy for gastric cancer, assess their relationships with patients’ 5-year overall survival (OS) and recurrence-free survival following surgery, and provide an immunological foundation for prognostic assessment.

METHODS

A total of 112 patients with stage I-III gastric cancer who underwent R0 resection between June 2018 and June 2020 were included retrospectively. Paraffin-embedded specimens of postoperative cancer tissues and adjacent tissues were collected. The multicolor immunohistochemistry technique was used, and pan-cytokeratin staining was performed synchronously with 4,6-diamidino-2-phenylindole nuclear staining. The sections were scanned by the Vectra Polaris™ full-spectrum imaging system, and cell phenotype identification and spatial localization analysis were performed via InForm^® 2.6 software. The independent prognostic value of OS/disease-free survival and immunological markers was assessed (tumor-node-metastasis stage, Lauren classification), and the Kaplan-Meier survival curve was used to analyze the survival differences between the high/low immune infiltration groups (log-rank test).

RESULTS

The density of CD8+ T cells in the invasion margin area (median: 248.5/mm²) was significantly greater than that in the tumor core area (108.3/mm², P < 0.001). The 5-year OS rate was high in patients with a high CD8+ density (> 220/mm²; 68.2%), low-density group: 42.1%, P = 0.003. An independent correlation between extended disease-free survival and a CD8+/forkhead box protein 3+ (FoxP3+) ratio > 2.5 was found [hazard ratio (HR) = 0.47, 95% confidence interval (CI): 0.29-0.76]. The risk of death was greater for those with a fraction of programmed death ligand 1+ tumor cells > 10% (HR = 2.15, 95%CI: 1.32-3.51). The enrichment of CD68+CD163+ M2 macrophages in the core area of the tumor predicted an increased risk of recurrence (HR = 1.72, P = 0.018). An immune prognosis scoring system was constructed on the basis of least absolute shrinkage and selection operator (regression, integrating the density of the CD8+ marginal zone, the CD8+/FoxP3+ ratio, and the 5-year OS prediction area under the curve was 0.81 (95%CI: 0.74-0.88).

CONCLUSION

The spatial distribution pattern of immune cells in the tumor microenvironment following gastric cancer surgery and the percentages of particular subgroups (programmed death ligand 1+ and CD8+/FoxP3+ tumor cells) are independent predictors of long-term survival.

Key Words: Correlation analysis; Immune cell infiltration; Tumor microenvironment; Gastric cancer; Over survival

Core Tip: This retrospective cohort study evaluated spatial distribution patterns of immune cells in the tumor microenvironment after radical gastrectomy for gastric cancer using multiplex immunohistochemistry. High CD8+ T-cell density at the invasive margin, a CD8+/forkhead box protein 3+ ratio > 2.5, and low programmed death ligand 1+ tumor cell proportion were independent predictors of favorable survival. An immune prognostic score integrating these features achieved high accuracy in predicting 5-year overall survival. These findings highlight the prognostic value of immune cell spatial profiling and may guide individualized postoperative immunotherapy strategies.

INTRODUCTION

The incidence and mortality rates of gastric cancer have decreased compared with those reported in previous statistics in 2022. The majority of cases of stomach cancer occur in underdeveloped nations, yet it is still one of the five most common cancers worldwide[1-4]. Moreover, under the influence of risk factors such as population aging, regional differences, dietary habits and Helicobacter pylori infection in our country, gastric cancer remains a malignant factor affecting the lifespan of people[5-7]. Early-stage gastric cancer has a low early diagnosis rate because it is often overlooked due to its lack of obvious clinical symptoms like gastritis and gastric ulcers, or because it often presents nebulous symptoms like upper abdominal discomfort and burp, which are similar to the symptoms of chronic ulcerative illnesses of the digestive system. The prognosis of patients with stomach cancer is significantly impacted by the fact that the majority of these patients receive their diagnosis at an advanced stage[8-10]. The current treatment approaches for gastric cancer have steadily changed into a comprehensive treatment model that combines surgical resection, postoperative radiotherapy and chemotherapy, neoadjuvant chemotherapy, molecular targeted therapy, and immunotherapy as a result of the ongoing advancements in medical technology[11]. Nonetheless, the mortality and recurrence rates of stomach cancer continue to be quite high[12]. Finding new targets for the prognostic evaluation of gastric cancer patients is therefore crucial.

In research on molecular markers related to prognosis, the number of reports on genetic models in the literature has also increased annually[13]. On the basis of the differential analysis of the tumor interstitial score and immune score of gastric cancer patients in The Cancer Genome Atlas database, a total of 45 genes with expression differences were obtained. After Cox regression analysis, a gastric cancer prognostic model including SRY-box transcription factor 9, leucine rich repeat containing 32, cat eye syndrome chromosome region, candidate 1 and membrane-spanning 4-domains A4A was established[14]. Changes in miRNA levels can be used to build new diagnostic and prognostic models based on variations in miRNA expression. Furthermore, numerous investigations have been carried out using gene expression levels in conjunction with pertinent analytical techniques to evaluate genes linked to prognosis and build risk ratio models[15,16]. According to recent research, malignant tumor cells are frequently hypoxic[17-19]. Hypoxia may be the cause of local cancers’ rapid metastasis or dissemination. Tumors that have inadequate or erratic blood perfusion frequently exhibit hypoxia characteristics. Certain transcription factors may become active in tumor cells in hypoxic conditions, triggering subsequent signals that control cell division, migration, and death. Research has demonstrated that hypoxia can trigger the growth, invasion, and spread of cancers, including hepatocellular carcinoma, colorectal cancer, and breast cancer[20-22]. In malignancies, hypoxia can also result in immunosuppression and immune evasion. Hypoxia causes T cells to become exhausted or nonresponsive, which results in functional problems of the body. In the human body, macrophages are essential. In order to treat a variety of illnesses, they are generated from monocytes in the blood system and go through a number of activation processes before developing into two distinct types of macrophages, M1 and M2. M1 macrophages are very good effector cells for defense. But M2 macrophages are better at inducing angiogenesis and tissue remodeling and are less effective at delivering antigens[23]. In recent years, the tumor microenvironment, a key factor in tumor occurrence, development and metastasis, has received extensive attention[24,25]. Immunohistochemical analysis of the tumor tissues of postoperative gastric cancer patients was quantitatively performed. Combined with clinical follow-up data, the relationships between these variables and the long-term survival of patients were analyzed. We expect that the findings of this study provide a scientific basis for the formulation of individualized immunotherapy strategies.

By elucidating their mode of action in tumor recurrence and metastasis, it is possible to guide more precise treatment and management methods in clinical practice by clarifying the infiltration characteristics of immune cells in the tumor microenvironment following gastric cancer surgery.

MATERIALS AND METHODS

Study design and patient cohort

This study is a retrospective cohort study. The inclusion criterion was patients with gastric cancer who underwent radical gastrectomy (R0 resection) between June 2018 and June 2020. The postoperative pathological stage was I-III. The exclusion criteria were as follows: Patients receiving neoadjuvant therapy (chemotherapy or chemoradiotherapy); patients who underwent nonradical surgery (R1 or R2 resection) after the operation; patients with incomplete clinical data or follow-up data; and patients who were diagnosed with residual gastric cancer by postoperative pathology. This research protocol was reviewed and approved by the Ethics Committee of Huai’an Hospital of Huai’an City (Approval No. HK20250610-12).

Clinical data collection and follow-up

Age, sex, tumor location, tumor size, histological type (as defined by the World Health Organization), Lauren classification (intestinal, diffuse, or mixed), T stage, N stage, M stage, final tumor-node-metastasis ( stage, surgical technique, degree of differentiation, postoperative adjuvant treatment plan (if any), and other baseline data were systematically gathered from the hospital’s electronic medical record system and pathological information system.

Survival follow-up: The primary study endpoints were overall survival (OS) and recurrence-free survival/disease-free survival (DFS). OS was defined as the time from the date of surgery to death for any reason.

Follow-up methods: Regular outpatient re-examination records, phone follow-ups, and inquiries into the resident death registration system are used to carry them out. The survival period of patients who were still alive and had not experienced a recurrence by the deadline was considered to be erased data.

Tissue specimen processing

The paraffin block archive of the hospital’s pathology department was used to get all paraffin-embedded tissue blocks from the surgical samples of the included patients who had undergone radical gastrectomy for stomach cancer. From each patient’s tumor and surrounding normal regions (greater than 3 cm from the cancer lesion’s margin), representative tissue blocks were chosen. A paraffin sectioning machine was used to constantly cut tissue slices that were 4 μm thick. Every segment was fastened to a glass slide that was positively charged.

Multiplex immunohistochemistry

An Opal™ 7-plex fluorescence immunohistochemical kit based on Tyramide signal amplification technology (NEL797001KT; PerkinElmer, Waltham, MA, United States). The following 7 target molecules were simultaneously labeled: Nuclear labeling: 4’,6-diamidino-2-phenylindole; tumor cell markers: Pan-cytokeratin and broad-spectrum cytokeratin; and immune cell markers.

Statistical analysis

All the statistical analyses were performed via SPSS software version 26.0 (IBM Corp., Armonk, NY, United States) and R software (Statistical Computing, Vienna, Austria). X-tile software, which is based on the principle of maximum selection rank statistics, was used for the training set or all sample data to determine the optimal cutoff value for each key immune indicator [CD8+ density, the CD8+/forkhead box protein 3+ (FoxP3+) ratio, the programmed death ligand 1 (PD-L1)+ ratio, and the CD68+CD163+ ratio] to predict 5-year OS or recurrence-free survival (CD8+ intratumoral microenvironment density > 220) cells/mm²; the ratio of CD8+ to FoxP3+ cells was greater than 2.5.

RESULTS

General clinical data

Numerous factors may have an impact on the prognosis, according to the clinical data and treatment-related information of 112 patients with stage I-III gastric cancer. Particularly in terms of OS and DFS, patients who underwent preoperative neoadjuvant chemotherapy and postoperative adjuvant chemotherapy had a better overall prognosis than those in the surgery group. Targeted therapy improves the prognosis for patients who test positive for human epidermal growth factor receptor 2. 5-year OS and DFS were higher in patients with high CD8+ T-cell density and PD-L1+.

Prolonged DFS and OS were independently linked to a low PD-L1+ tumor cell ratio and a high CD8+/FoxP3+ ratio. Recurrence risk is substantially elevated in correlation with high CD68+CD163+ M2 macrophage infiltration (Table 1). In addition, only six variables were included in the prognostic index: T stage, N stage, and M stage. Age [hazard ratio (HR) = 2.158, 95% confidence interval (CI): 1.377-3.383, P = 0.001), prognostic index (HR = 1.606, 95%CI: 1.088-2.369, P = 0.017), and M stage (HR = 2.322, 95%CI: 1.175-4.587, P = 0.015) were found to be independent factors influencing the prognosis of gastric cancer patients. The age HR was 2.158, suggesting that the prognosis of patients over 60 years old was poorer than that of patients under 60 years old. The HR of the prognostic index was 1.606. The HR of the M stage was 2.322, suggesting that the prognosis of patients with tumors without distant metastasis was better than that of patients with distant metastasis (Table 2).

Table 1 Clinical data and treatment statistics of gastric cancer patients.

Feature	Quantity (n = 112)	Percentage (%)	Remarks
Age	59
Gender
Male	68	60.7
Female	44	39.3
TNM installment
I	25	22.3
II	48	42.9
III	39	34.8
Lauren classification
Intestinal type	63	56.2
Diffuse type	49	43.8
Histological grading
Highly differentiated	23	20.5
Moderate differentiation	53	47.3
Low differentiation	36	32.2
Treatment plan
Only surgery	34	30.4
Preoperative neoadjuvant chemotherapy	19	17	Including FOLFOX and XELOX solutions
Postoperative adjuvant chemotherapy	59	52.7	Including FOLFOX and XELOX solutions
Marker expression
HER2 positive	21	18.8
HER2 negative	91	81.2
PD-L1 expression
PD-L1 positive (≥ 1%)	43	38.4
PD-L1 negative (< 1%)	69	61.6
MSI status
MSI-H	12	10.7
MSS	100	89.3
Immune cell infiltration
High density of CD8+ T cells	54	48.2	> 220/mm2
Low density of CD8+ T cells	58	51.8	≤ 220/mm2
High CD8+/FoxP3+ ratio	61	54.5	> 2.5
Low CD8+/FoxP3+ ratio	51	45.5	≤ 2.5
High proportion of PD-L1+ tumor cells	22	19.6	> 10%
Low proportion of PD-L1+ tumor cells	90	80.4	≤ 10%
Macrophages with high CD68+CD163+ M2	49	43.8
Low CD68+CD163+ M2 macrophages	63	56.2
Follow-up results
5-year OS rate	57	50.9
5-year DFS rate	65	58
Recurrence situation

TNM: Tumor-node-metastasis; HER2: Human epidermal growth factor receptor 2; PD-L1: Programmed death ligand 1; MSI: Microsatellite instability; MSI-H: Microsatellite instability-high; MSS: Microsatellite stability; FoxP3+: Forkhead box protein 3+; OS: Overall survival; DFS: Disease-free survival; FOLFOX: Folinic acid, fluorouracil, and oxaliplatin; XELOX: Capecitabine and oxaliplatin.

Table 2 Multivariate study of clinical variables using regression.

Clinical indicators	Univariate Cox	P value	Multivariate Cox	P value
Age (≤ 60, > 60)	99.12	0.002	21.58	0.001
Gender (male, female)	0.674 (0.445-1.021)	0.062
Stage (1-2, 3-4)	1.737 (1.170-2.578)	0.006	1.155 (0.560-2.384)	0.697
T (1, 2, 3, 4)	1.298 (1.023-1.645)	0.032		0.564
N (1, 2, 3)	1.267 (1.069-1.502)	0.006		0.345
M (0, 1)	2.048 (1.096-3.827)	0.025	2.322 (1.175-4.587)	0.015
PI (< 82.506945, ≥ 82.506945)	1.753 (1.199-2.563)	0.004	1.606 (1.088-2.369)	0.017

T: Tumor; N: Node; M: Metastasis; PI: Prognostic index.

Spatial distribution characteristics of immune cells

Important immune cells were shown to exhibit notable spatial variability in the tumor microenvironment following gastric cancer surgery, according to multicolor immunohistochemistry and spatial localization analysis. The tumor core had a higher enrichment degree than the invasion margin (tumor core: 84.7 cells/mm²vs invasive margin: 52.4 cells/mm², P = 0.006), and the infiltration density at the tumor’s invasive margin (median: 248.5 cells/mm²) was significantly higher than that in the tumor core (108.3 cells/mm², P < 0.001; Wilcoxon test; Table 3).

Table 3 Immune cell number and geographic distribution in the tumor microenvironment, median (interquartile range).

Immune cell phenotype	Core area of the tumor (cells/mm2)	Invasion edge area (pieces/mm2)	P value
CD8+ T cells	108.3 (75-142)	248.5 (195-312)	< 0.001
FoxP3+ Treg	42.1 (28-58)	48.9 (34-67)	0.215
CD68+CD163+ M2	84.7 (62-113)	52.4 (39-74)	0.006
CD4+ T cells	112.3 (79-152)	262.5 (199-344)	< 0.001
Treg cells	89.7 (72-123)	72.4 (42-84)	< 0.001
NK cells	52.1 (33-59)	55.9 (44-69)	< 0.001
M1 cells	133.3 (84-159)	277.5 (212-364)	< 0.001
M2 cells	122.3 (86-162)	282.5 (179-394)	< 0.001

FoxP3+: Forkhead box protein 3+.

Analysis of positive correlation factors for prognosis

Compared to the tumor core area (108.3/mm²), the infiltration margin area had a considerably higher median density of CD8+ T cells (248.5/mm²; P < 0.001). Compared to patients in the low-density group, those with a high CD8+ T-cell density (> 220/mm²) had a 5-year OS rate of 68.2%, which was 42.1%, P = 0.003. Recurrence-free survival was considerably longer in patients with a CD8+/FoxP3+ ratio higher than 2.5 (HR = 0.47, 95%CI: 0.29-0.76). A higher CD8+/FoxP3+ ratio means that there are more antitumor effector cells (CD8+ T cells) in the tumor microenvironment than immunosuppressive cells (FoxP3+ Treg cells), which helps stop tumors from coming back. On the basis of least absolute shrinkage and selection operator regression analysis, we constructed an immune prognosis scoring system that integrates the density of CD8+ marginal areas and the proportions of PD-L1+ tumor cells and CD8+/FoxP3+ cells. The area under the curve (AUC) of this scoring system for predicting the 5-year OS rate reached 0.81 (95%CI: 0.74-0.88), demonstrating high predictive accuracy (Table 4).

Table 4 Analysis of positive correlation factors for prognosis.

Variable	Median/proportion	Five-year overall survival rate	P value
CD8+ T-cell density	248.5/mm2	High-density group 68.2% vs low-density group 42.1%	0.003
The ratio of CD8+ to FoxP3+	> 2.5
PD-L1+ tumor cell ratio	> 10%	Increase the risk of death
CD68+CD163+ M2 macrophages		Increase the risk of recurrence
IPS scoring system		AUC = 0.81

FoxP3+: Forkhead box protein 3+; PD-L1: Programmed death ligand 1; IPS: Immune prognosis scoring system.

Correlation analysis of the risk score with immune cell and immune infiltration scores

There is a correlation between risk difference genes and the activation and function of immune cells. The results revealed that the risk score was correlated with the infiltration of dendritic cells, macrophages and neutrophils. Moreover, the higher the risk value is, the higher the cell content. Sangerbox online software was subsequently used to conduct an ESTIMATEScore analysis of the tumor microenvironment, and the obtained results were plotted and visualized. The abundance of stromal cells in the high-risk group was greater than that in the low-risk group. Moreover, stromal cells are considered by most studies to play important roles in tumor growth, disease progression and drug resistance. These findings indicate that abnormal immune infiltration in gastric cancer patients may be a prognostic indicator and target of immunotherapy and may have important clinical significance (Figures 1 and 2).

Figure 1 High-risk and low-risk groups and immune infiltration of various immune cells. Correlation analysis between risk score and immune cell infiltration levels. Scatter plots displaying the correlations between risk score (X-axis) and the abundance of six immune cell types (Y-axis). CD8+ T cells (correlation coefficient = 0.411, P = 3.052 × 10^-18); CD4+ T cells (correlation coefficient = 0.587, P = 6.238 × 10^-40), cancer-associated fibroblasts (correlation coefficient = 0.683, P = 3.354 × 10^-58), dendritic cells (correlation coefficient = 0.302, P = 3.455 × 10^-10), tumor-associated macrophages (correlation coefficient = 0.453, P = 5.05 × 10^-22), and neutrophils (correlation coefficient = 0.218, P = × 10^-5). All immune cell types showed significant positive correlations with risk score, with cancer-associated fibroblasts demonstrating the strongest correlation. Each dot represents an individual sample, and the gray shaded area indicates the 95% confidence interval of the regression line. Cor: Correlation coefficient.

Figure 2 High-risk and low-risk groups and immune infiltration of various immune cells. Comparison of tumor microenvironment scores between high-risk and low-risk groups. Violin plots showing the distribution differences between the high-risk group (red) and low-risk group (green) across three scoring metrics: StromalScore, ImmuneScore, and ESTIMATEScore. All three scores showed statistically significant differences between the two groups (P < 0.0001), with the high-risk group exhibiting significantly higher stromal infiltration, immune infiltration, and overall tumor microenvironment scores compared to the low-risk group. The violin shapes display the distribution density of the data, with internal scatter points representing the actual score values for individual samples.

Correlation analysis between a single immune indicator and survival

The critical value of CD8+ T cells at the invasion margin (220/mm²) was determined according to the receiver operating characteristic (ROC) curve. The risk of death was considerably greater for patients with a fraction of PD-L1+ tumor cells > 10% (HR = 2.15, 95%CI: 1.32-3.51; P = 0.001), and the 5-year OS rate was only 36.8% (vs 62.4% in the low-expression group). Patients who had more than 85/mm² CD68+CD163+ M2 cells in the tumor core had a 72% greater chance of recurrence (HR = 1.72, 95%CI: 1.10-2.70, P = 0.018), which was significantly associated with early metastasis (< 24 months; Tables 5 and 6).

Table 5 Key immune indicators for survival prognosis using multivariate Cox regression analysis.

Variable	OS	DFS	P value
	HR (95%CI)	HR (95%CI)
CD8+ IM density > 220/mm2	0.52 (0.33-0.82)	0.61 (0.40-0.93)	0.004
CD8+/FoxP3+ ratio > 2.5	0.69 (0.45-1.06)	0.47 (0.29-0.76)	0.09
PD-L1+ tumor cells > 10%	2.15 (1.32-3.51)	1.84 (1.20-2.82)	0.001
CD68+CD163+ M2 TC > 85/mm2	1.43 (0.92-2.22)	1.72 (1.10-2.70)	0.112

IM: Invasive margin; FoxP3+: Forkhead box protein 3+; PD-L1: Programmed death ligand 1; TC: Tumor core; OS: Overall survival; DFS: Disease-free survival; HR: Hazard ratio; CI: Confidence interval.

Table 6 Independent prognosis of immune indicators evaluated by Cox proportional hazards model.

Variable	HR	95%CI	P value
PD-L1+ tumor cell ratio	2.15	1.32-3.51	0.002
CD68+CD163+ M2 macrophages (tumor core area)	1.72	1.10-2.68	0.018
CD8+ T-cell density (invasion margin area)	0.63	0.45-0.86	0.004
CD8+/FoxP3+ proportion	0.47	0.29-0.76	0.002
Age	1.03	1.01-1.05	0.011

PD-L1: Programmed death ligand 1; FoxP3+: Forkhead box protein 3+; HR: Hazard ratio; CI: Confidence interval.

Analysis of clinicopathological factors of gastric cancer

Univariate Cox analysis revealed high expression of biglycan. There was a strong correlation between the OS of gastric cancer patients and the region of gastric cancer occurrence, tumor size, vascular tumor thrombus, and nerve invasion (Table 7). High biglycan expression, tumor size, depth of invasion, lymph node metastasis, clinical stage, and vascular tumor thrombus were identified via multivariate Cox regression analysis. In gastric cancer patients, nerve invasion and other parameters were found to be independent risk factors for OS (P < 0.05, Table 8).

Table 7 Univariate Cox analysis of clinicopathological factors and prognosis in patients with gastric cancer.

Clinicopathological factors	HR	95%CI	P value
Age
< 65 ages	1.000
≥ 65 ages	1.306	0.720-2.369	0.379
Gender
Male	1.000
Female	0.860	0.467-1.584	0.628
Tumor location
Lower district	1.000
Central district	1.275	1.067-1.525	< 0.05
Whole stomach	4.753	3.304-5.181	< 0.05
Upper district	1.260	1.051-1.510	< 0.05
Tumor size
< 5.0 cm	1.000
≥ 5 cm	2.370	1.342-4.185	0.003
Depth of infiltration
T1	1.000
T2	1.863	0.944-3.196	< 0.05
T3	2.044	0.340-12.293	< 0.05
T4	2.257	0.694-7.346	< 0.05
Lymph node metastasis
N0	1.000
N1	2.405	1.705-3.392	< 0.05
N2	3.487	1.593-7.630	0.002
N3	6.072	2.554-14.433	< 0.001
Distant transfer
M0	1.000
M1	2.196	0.854-5.646	0.103
Pathological staging
I	1.000
II	2.024	1.430-5.826	< 0.05
III	3.740	1.435-9.747	0.007
IV
Vascular tumor thrombus
Negative	1.000
Positive	1.880	1.684-2.200	0.039
Nerve infiltration
Negative	1.000
Positive	2.666	1.119-6.349	0.027
BGN expression score
< 4 points	1.000
≥ 4 points	4.123	1.234-7.748	0.001

T: Tumor; N: Node; M: Metastasis; BGN: Biglycan; HR: Hazard ratio; CI: Confidence interval.

Table 8 Multivariate Cox analysis of the prognosis and clinicopathological features of individuals with stomach cancer.

Factor	HR	95%CI	P value
Tumor location	1.165	1.036-1.311	> 0.05
Tumor size	1.565	1.286-1.904	< 0.05
Depth of infiltration	1.656	0.944-3.196	< 0.05
Lymph node metastasis	3.487	1.593-7.630	< 0.05
Distant transfer	2.196	0.854-5.646	< 0.05
Pathological staging	3.740	1.435-9.747	< 0.05

HR: Hazard ratio; CI: Confidence interval.

Determination of the optimal cutoff values for each immune indicator

The median value of the CD4+/CD8+ ratio was 1.16 (0.52-1.94), and the median values of Treg cells and natural killer cells were 9.4% (6.43%-19.64%) and 17.3% (7.52%-23.78%), respectively. The ROC curve was plotted with the OS of the patients as the endpoint. ROC curve analysis revealed that the optimal cutoff values for the CD4+/CD8+ T-cell ratio and the Treg and natural killer-cell counts were 1.39%, 7.8% and 19.9%, respectively (Figure 3).

Figure 3 Receiver operating characteristic curve analysis of immune indicators and overall survival. Receiver operating characteristic curves evaluating the predictive performance of immune cell infiltration levels. Three receiver operating characteristic curves are shown for CD4+/CD8+ T cell ratio, Treg cells, and natural killer cells in predicting patient outcomes or risk stratification. The diagonal reference line represents random chance (area under the curve = 0.5). The curves demonstrate the sensitivity (true positive rate, Y-axis) vs specificity (false positive rate, X-axis) across different threshold values. The deviation of each curve from the diagonal line indicates the discriminatory power of the respective immune cell marker in the prediction model. NK: Natural killer.

DISCUSSION

Systematic examination of the infiltration patterns of immune cells in the surrounding tissue of the tumor following gastric cancer surgery and their connection to patients’ long-term survival revealed the complex involvement of immune cells in the tumor microenvironment[26]. This discovery provides a new perspective for understanding the impact of the tumor microenvironment after gastric cancer surgery on disease progression and patient prognosis. This study focused on immune cell infiltration patterns in the tumor microenvironment and how these patterns are related to patient prognosis after gastric cancer surgery[27]. By thoroughly analyzing the infiltration patterns of different immune cells, we not only revealed the role of immune cells in antitumor immunity but also identified potential targets for future immunotherapy of gastric cancer[28-31].

The density of CD8+ T cells in the infiltration margin area (median: 248.5/mm²) was significantly greater than that in the tumor core area (108.3/mm², P < 0.001). This result indicates that the high-density distribution of CD8+ T cells at the tumor invasion margin may play an important role in combating the tumor immune response. Further analysis revealed that patients with a CD8+ T-cell density greater than 220/mm² had a 5-year OS rate of 68.2%, which was significantly greater than the 42.1% reported in the low-density group (P = 0.003). A higher CD8+/FoxP3+ ratio may represent a more effective antitumor immune response, thereby delaying tumor recurrence. Furthermore, for patients whose proportion of PD-L1+ tumor cells exceeded 10%, the risk of death significantly increased (HR = 2.15, 95%CI: 1.32-3.51), indicating that tumors with high PD-L1 expression levels may reduce the survival rate of patients through immune escape mechanisms. CD68+CD163+ M2-type macrophages enriched in the core area of the tumor predicted a greater risk of recurrence (HR = 1.72, P = 0.018), suggesting that M2-type macrophages may play a protumor role in tumor progression. The immune prognosis scoring system constructed on the basis of least absolute shrinkage and selection operator regression integrates the density of the CD8+ marginal zone and the ratio of CD8+/FoxP3+ cells. The AUC for predicting 5-year OS reached 0.81 (95%CI: 0.74-0.88), demonstrating the excellent performance of this system in prognosis assessment.

It is possible to assess the prognosis of patients more precisely by identifying the pattern of immune cell infiltration in the postoperative tumor microenvironment, thereby guiding postoperative follow-up and the formulation of adjuvant treatment strategies[32-34]. The analysis of the infiltration patterns of different types of immune cells provides new ideas for immunotherapy[35]. For example, immunotherapy strategies that target T-cell activation and function enhancement or therapeutic methods that regulate the polarization state of macrophages are expected to improve treatment outcomes for patients with gastric cancer[36-38]. In addition, the findings of this study suggest that the combined application of immune checkpoint inhibitors and other immunomodulators may produce a synergistic effect, further improving the prognosis of patients.

CONCLUSION

The link in the tumor microenvironment following gastric cancer surgery is shown in this work, which offers a new avenue for research and a practical application strategy that will greatly improve treatment efficacy and patient survival. To validate and advance the results of this investigation, large-scale, multicenter clinical trials must be conducted.