Introduction

Early gastric cancer (EGC) is defined by cancer cells invading the mucosa (T1a) or submucosa (T1b) of the gastric wall, regardless of regional lymph node metastasis (LNM). The advent of widespread endoscopic use has facilitated increased EGC detection. Data from the Chinese Surgical Alliance for Gastrointestinal Oncology indicates that EGC accounted for 19.7% of all gastric cancer cases, rising to 20.9% between 2014 and 2017.¹ After radical resection, the 5-year overall survival (OS) rate for EGC can surpass 90%, yet recurrence and metastasis still occur in 1.1%–13.8% of patients.^2–5 Hence, accurate prognostic assessment of individual outcomes is crucial for guiding treatment strategies and follow-up protocols in managing gastric cancer. In EGC, where tumor invasion is confined to the mucosal layer, special attention should be paid to LNM. Studies have shown that the 5-year OS rate of EGC patients with the N1 stage of LNM is significantly higher than that of patients with N2 and N3 stages of LNM.⁶

The Tumor, Node, Metastasis (TNM) Staging System is widely used in clinical practice to assess the biological characteristics of tumors. However, there exists a complex interplay between the tumor and the host involving inflammatory responses and nutrient metabolism. Recent investigations have underscored the prognostic significance of inflammatory markers, derived from peripheral blood counts, across diverse malignancies.^7–10 Notably, NLR and PLR are extensively investigated prognostic indicators in gastric cancer.^11–13 The SII, integrating NLR and PLR, provides a comprehensive assessment of systemic inflammatory status and holds potential for predicting both tumor response and prognosis in patients with gastric cancer.¹⁴ Concurrently, nutritional parameters such as albumin, body mass index (BMI), and prognostic nutritional index (PNI) are also linked to prognosis in gastric cancer patients following gastrectomy.^15–17 This study aims to delve into the factors influencing prognosis among EGC patients post-radical surgery, encompassing immunoinflammatory and nutritional parameters alongside LNM.

Materials and Methods

Patient Sample

We retrospectively analyzed the clinicopathologic and follow-up data of 490 patients who were diagnosed with EGC and underwent gastrectomy and lymph node dissection at the Affiliated Hospital of Qingdao University between January 2013 and December 2018. The inclusion criteria of patients were as follows: (1) patients who underwent radical gastrectomy with R0 resection; (2) patients who received D2 lymph node dissection; (3) patients who had EGC limited to submucosal infiltration (pT1); (4) patients with complete clinicopathologic and follow-up data available. The exclusion criteria were as follows: (1) patients with distant metastases and/or other malignancies; (2) patients who received neoadjuvant chemotherapy and/or radiotherapy; (3) patients with R1 or R2 surgical margins; (4) patients with a history of autoimmune, inflammatory, or hematological disorders; (5) patients who received blood transfusions and nutritional supplements within one month before blood collection; (6) patients who died within 30 days after surgery.

Data Collection

A total of 490 patients meeting the inclusion / exclusion criteria were enrolled in this study. Postoperative outpatient review records, routine physical examination records, and yearly routine telephone follow-up were used to monitor the patients and gather data. The collected clinicopathological characteristics included age, sex, BMI, adjuvant chemotherapy, tumor location, gross type, lesion size, depth of invasion, Lymphovascular invasion (LVI), and perineural invasion. All patients underwent blood sample collection for routine laboratory tests within 7 days before surgery, with no disclosure of personal information during data collection. According to guidelines from the Japanese Gastric Cancer Association (JGCA), tumor locations were categorized as upper, middle, and lower stomach regions.¹⁸ The two classifications of gross types encompass elevated types, including EGC gross types I and IIa, and non-elevated types, which comprise EGC gross types IIb, IIc, and III.¹⁹ Depth of invasion and LNM status were assessed using the TNM staging system of the American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) for gastric cancer. Patients, with a mean age of 58.90 years, were stratified into ≤ 60 years and > 60 years age groups using a cutoff value of 60 years. Nutritional indicators such as the PNI and BMI were calculated as follows: PNI = albumin (g/L) + 5 × lymphocyte count (10⁹/L), BMI = weight (kg) / height squared (m²). Patients were categorized into non-obese (BMI < 28 kg/m²) and obese (BMI ≥ 28 kg/m²) groups based on obesity criteria. Inflammation-based indices including NLR, PLR, and SII were calculated as follows: NLR = neutrophils (N) / lymphocytes (L); PLR = platelets (P) / L; SII = P × N / L.

Statistical Analysis

Statistical analyses were performed using SPSS statistical software for Windows, version 26.0 (IBM, Chicago, Illinois, USA). Receiver operating characteristic (ROC) curves were utilized to determine optimal cutoff values for inflammatory and nutritional indices. Continuous variables were assessed via t-tests, while categorical variables were analyzed using chi-square tests or Fisher’s exact tests. Survival disparities, such as 5-year disease-free survival (DFS) and OS, were evaluated using Kaplan-Meier analysis with Log rank tests. Additionally, multifactorial Cox proportional hazards modeling was applied to identify independent prognostic factors influencing 5-year DFS in EGC patients. Statistical significance was set at a two-sided P value < 0.05. The area under the ROCcurve (AUC) was employed to gauge the prognostic accuracy of each parameter for 5-year DFS. The goodness of fit was assessed using the Hosmer-Lemeshow test, where a P value > 0.05 indicated a well-fitted model.²⁰ Decision curve analysis was conducted using R software (version 4.1.3) and various related packages to ascertain predicted net benefit thresholds and clinical effectiveness rates.²¹

Results

Baseline Characteristics

In this study, 490 patients diagnosed with EGC were analyzed, comprising 356 males and 134 females. The average age of the cohort was 58.90 years (range: 24 to 87 years), with 243 patients aged 60 years or older and 247 aged younger than 60 years. 409 patients (83.5%) had a BMI below 28 kg/m², while 81 cases (16.5%) had a BMI over or equal to 28 kg/m². The majority of tumors were located in the lower stomach (356 instances, 72.6%), followed by the middle (120 cases, 24.5%) and upper parts (14 cases, 2.9%). There were 394 cases (80.4%) of non-elevated type and 96 cases (19.6%) of elevated type. The depth of tumor invasion involved the mucosal layer in 229 cases (46.7%) and the submucosal layer in 261 cases (53.3%). 368 cases (78.8%) had a negative lymph node status, while 104 cases (21.2%) had metastases, including N1 (14.7%), N2 (4.9%), and N3 (1.6%). Based on lesion size, 298 patients (60.8%) had lesions ≤20 mm, 104 patients (21.2%) had lesions ≤30 mm, and 88 patients (18.0%) had lesions >30 mm. LVI was positive in 78 cases (15.9%) and negative in 412 cases (84.1%). Additionally, perineural invasion was present in 42 cases (8.6%), while 448 cases (91.4%) showed no invasive perineural elements. According to treatment, 151 patients (30.8%) had adjuvant chemotherapy after surgery, while 339 patients (69.2%) only had surgical resection. The specific information was displayed in Table 1.

Table 1 Baseline Characteristics of 490 Patients with Early Gastric Cancer Post-Surgery

The Optimal Cutoff Values for Inflammatory Index and Nutritional Index

ROC curve analysis was performed to ascertain the optimal cutoff values for each inflammatory and nutritional index, including NLR, PLR, SII, and PNI, which were identified as 1.88, 136.90, 613.05, and 42.21, respectively. Subsequently, patients were stratified into low and high groups according to these cutoff values. Values below 1.88 for NLR, 136.90 for PLR, 613.05 for SII, and 42.21 for PNI are categorized as the low-level group, while values equal to or above these thresholds are classified as the high-level group.

The Relationship Between SII, PNI, and Clinicopathologic Features

Among the 490 patients in the study, 401 (81.8%) were categorized into the low SII (SII-L) group, while 89 (18.2%) were placed in the high SII (SII-H) group. Analysis of clinicopathologic features revealed that the SII-H group exhibited a higher percentage of patients with tumor infiltration reaching the submucosal layer compared to the SII-L group (65.2% vs 50.6%, P = 0.013). Moreover, a greater occurrence of LNM was observed in the SII-H group (29.2% vs 19.5%, P = 0.042). However, no statistically significant differences were observed in other baseline characteristics such as sex, age, BMI, adjuvant chemotherapy, tumor location, lesion size, LVI, perineural invasion, and pathological N-staging between the SII-L and SII-H groups. Details were shown in Table 2.

Table 2 The Relationship Between SII, PNI, and Clinicopathologic Features

In this study, there were 455 patients (92.9%) in the low PNI (PNI-L) group and 35 patients (7.1%) in the high PNI (PNI-H) group. A higher percentage of patients with elevated type was observed in the PNI-H group compared to the PNI-L group (37.1% vs 18.2%, P=0.007; Table 2). Nevertheless, no significant differences were observed between the two groups in terms of other baseline characteristics.

Risk Factors for Lymph Node Metastasis

In the univariate analysis, variables such as gross type, lesion size, depth of invasion, LVI, perineural invasion, and SII were associated with LNM (Table 3). Subsequent multivariate logistic regression revealed that only depth of invasion and LVI remained independent predictors of LNM. Notably, LVI exhibited the highest odds ratio among these predictors (Table 4).

Table 3 Lymph Node Metastasis Risk According to Clinicopathologic Characteristics

Table 4 Multivariate Logistic Regression Analysis of Lymph Node Metastasis

Analysis of Prognostic Factors in Patients with Early Gastric Cancer

In this study, patients diagnosed with EGC demonstrated a favorable prognosis, achieving a 5-year DFS of 91.8% and a 5-year OS of 93.7%. In terms of survival analysis, the 5-year DFS and OS rates were significantly higher in the SII-L group at 94.7% and 96.2%, respectively, compared to the SII-H group, which had rates of 78.7% and 81.9%, respectively (P < 0.001 for both; Figure 1). Similarly, patients in the PNI-H group exhibited superior 5-year DFS and OS rates of 93.3% and 95.1%, respectively, compared to the PNI-L group, which showed rates of 71.4% and 74.3%, respectively (P < 0.001 for both; Figure 1).

Figure 1 Kaplan-Meier curves illustrating 5-year Disease-Free Survival (DFS) and Overall Survival (OS) comparisons for Prognostic Nutritional Index (PNI) (panels A and B) and Systemic Immune-Inflammation Index (SII) (panels C and D).

Univariate and multivariate analyses were conducted to identify independent prognostic risk factors for EGC recurrence. In univariate analysis, factors such as age (P = 0.003), depth of invasion (P = 0.003), pathological N-staging (P = 0.001), NLR (P < 0.001), PLR (P = 0.001), SII (P < 0.001), and PNI (P < 0.001) significantly influenced 5-year DFS in EGC patients. Conversely, variables such as sex, BMI, gross type, tumor location, adjuvant chemotherapy, lesion size, LVI, and perineural invasion showed no association with tumor recurrence (details in Table 5).

Table 5 Univariate Analysis of Prognostic Risk Factors for Early Gastric Cancer

Multivariate analysis revealed that LNM (P = 0.025), SII (P = 0.030), and PNI (P < 0.001) independently predicted increased risk of postoperative recurrence in EGC patients (Figure 2).

Figure 2 Forest plot depicting multivariate Cox proportional hazards regression analysis for postoperative recurrence in patients diagnosed with early gastric cancer.

Predictive Ability of SII and PNI

ROC curve analysis was employed to evaluate the prognostic efficacy of preoperative SII and PNI in EGC. SII demonstrated superior predictive ability (AUC = 0.637, 95% CI 0.551–0.723, P = 0.001) compared to PNI (AUC = 0.595, 95% CI 0.507–0.684, P = 0.022) and LNM stage (AUC = 0.611, 95% CI 0.525–0.697, P < 0.001) (Figure 3). Combined analysis of SII, PNI, and LNM for integrated prediction yielded a sensitivity of 0.704, a specificity of 0.697, and a C-index of 0.723 (P = 0.008) (Table 6). The Hosmer-Lemeshow test indicated a good model fit (P = 0.200). The inclusion of SII and PNI significantly enhanced the predictive accuracy of 5-year DFS, detailed in Table 6. Decision curve analysis demonstrated that the combination of SII and PNI with lymph node status for prognostic assessment had high clinical efficacy and practicality (Figure 4).

Table 6 Predictive Value of SII, PNI, Pathological N-Staging, and Their Combination for Prognosis in Early Gastric Cancer

Figure 3 Receiver operating characteristic (ROC) curves evaluating the predictive value of preoperative SII, PNI, N stage of lymph node metastasis, and their combined use for postoperative prognosis in gastric cancer patients.

Figure 4 Decision curve analysis. The y-axis indicates net benefit, calculated as true benefit (gain) minus false benefit (harm). The gray line denotes the assumption where all patients progress or perish, while black horizontal lines signify the hypothesis of no patient progression or mortality. The x-axis represents threshold probability.

Subgroup Analysis of the Prognostic Value of SII and PNI

In our subgroup analysis stratified by age, patients were categorized into non-elderly (< 60 years) and elderly (≥ 60 years) groups. Among elderly patients, the 5-year postoperative DFS rate was significantly lower in the PNI-L group compared to the PNI-H group (57.1% vs 91.3%, P < 0.001). Conversely, in the non-elderly group, there was no statistically significant difference observed in the 5-year DFS rates between the PNI-L and PNI-H groups (95.2% vs 92.9%, P = 0.280). Furthermore, within the elderly group, the 5-year DFS rate was markedly higher in the SII-L group compared to the SII-H group (92.6% vs 72.0%, P < 0.001). Similarly, in the non-elderly group, the 5-year DFS rate favored the SII-L group over the SII-H group (96.6% vs 87.2%, P = 0.004). Please refer to Figure 5 for additional details.

Figure 5 5-year DFS curves comparing PNI-H group versus PNI-L group stratified by age: (A) Elderly group and (B) Non-elderly group. 5-year DFS survival curves comparing high SII-H group versus SII-L groups: (C) Elderly group and (D) Non-elderly group.

Discussion

With the advancement of medical technology and the widespread use of endoscopy, an increasing number of gastric cancers are being detected in the early stages. According to data from the Chinese Surgical Alliance for Gastrointestinal Oncology, the proportion of EGC rose from 19.7% to 20.9% between 2014 and 2017.¹ The tumor, lymph node, metastasis (TNM) staging system, developed by the American Joint Committee on Cancer (AJCC) and the International Union for Cancer Control (UICC), is widely employed by clinicians to predict prognosis and guide treatment decisions. However, the TNM staging system alone incompletely accounts for variations in patient survival outcomes. It primarily describes the tumor’s growth and metastatic potential, overlooking the intricate interplay between the tumor and the body’s environment. Therefore, there is a need for further investigation into better prognostic indicators for patients with EGC to complement TNM staging. To date, various inflammatory markers in peripheral blood, such as NLR, PLR, and SII, as well as nutritional status indicators like PNI, have been recognized as useful prognostic biomarkers for cancers, including gastric cancer.^11,12,14,15 Our study aimed to investigate the clinical significance of peripheral blood inflammation and nutrition-related parameters (NLR, PLR, SII, PNI), along with clinicopathologic factors, in predicting survival outcomes for early gastric cancer.

Systemic inflammatory responses are closely linked to tumor development, especially in gastric cancer. Before cancer develops, there is a series of histological changes, beginning with persistent chronic active inflammation, followed by steps such as intestinal epithelial chemotaxis and heterogeneous hyperplasia that ultimately progress to cancer.²² Neutrophils, platelets, and lymphocytes are integral to systemic inflammation and immune responses, crucially influencing cancer development.

Tumor-associated neutrophils are a heterogeneous population characterized by plasticity, capable of displaying either anti-tumor (N1 type) or pro-tumor (N2 type) activities.²³ The tumor microenvironment (TME) contains various cytokines, chemokines, and signals that can recruit. Tumor-associated neutrophils and polarize them into a pro-tumor phenotype once they enter the TME.²⁴ This phenotypic shift can facilitate epithelial-mesenchymal transition (EMT) in gastric cancer cells through cytokine secretion,²⁵ stimulate tumor angiogenesis,^26,27 mediate immunosuppression and immune tolerance,^28,29 and significantly influence tumor cell migration,³⁰ invasion, and metastasis.³¹

Cancer patients often exhibit hypercoagulable blood, and cancer-related thrombocythemia is linked to adverse clinical outcomes.³² Platelets from gastric cancer patients can regulate the expression of genes involved in EMT, like matrix metalloproteinase 9 (MMP-9). This modulation subsequently increases malignant behaviors such as cancer cell proliferation, migration, invasion, and adhesion.³³ Platelets can also inhibit T cell immunity through the GARP-TGFβ axis,³⁴ contributing to cancer immune escape.³⁵ Lymphocytes, as a fundamental component of cytotoxic immune response, play a critical role in inhibiting tumor cell growth and proliferation, as well as in killing tumor cells.³⁶

Due to the rapid growth of the tumor, which competes for nutrients and occupies the gastrointestinal tract, along with patients experiencing gastrointestinal symptoms and other discomforts, cancer patients are susceptible to malnutrition and chronic wasting. Serum albumin, synthesized by the liver, is often used as a marker to reflect the nutritional status of the body. It is regulated by pro-inflammatory cytokines, which can be produced by both the host and the tumor and play a pivotal role in cancer progression and neovascularization.³⁷

Therefore, a comprehensive hematologic profile incorporating these variables may serve as a prognostic indicator for survival in cancer patients.

In our investigation, we comprehensively evaluated the role of preoperative PLR, NLR, SII, PNI, and other clinicopathologic indicators in predicting LNM in EGC. Our findings revealed that PLR, NLR, SII, and PNI did not exhibit significant correlations with the occurrence of LNM. Nonetheless, previous research has suggested that PLR may possess significant predictive efficacy for LNM in EGC.³⁸ The predictive roles of NLR and PLR in LNM among EGC patients remain subjects of ongoing debate.³⁹

In our multifactorial logistic analysis, we identified that depth of invasion and LVI are independent risk factors for LNM in EGC, with particularly high odds ratios observed for LVI. Studies suggest that submucosal invasion strongly correlates with LNM in EGC,⁴⁰ with lower rates observed in low-grade T1a compared to T1b tumors,⁴¹ attributable to the submucosa’s rich vascular and lymphatic network facilitating cancer cell dissemination. LVI positivity emerges as the most robust predictor of LNM in EGC.⁴² Furthermore, as LNM increases, the occurrence of LVI also elevates.⁴³ Importantly, patients affected by LVI and LNM experience diminished survival outcomes.⁴⁴ For patients with EGC showing LNM and LVI, targeted postoperative adjuvant chemotherapy significantly enhances the prognosis.⁴⁵

In recent years, the SII has garnered attention as a potentially superior prognostic marker compared to other inflammatory indicators, yet its precise prognostic value in EGC remains unclear. Our study focused on patients with pT1Nx EGC, revealing both the prognostic significance of the prognostic nutritional PNI and SII as independent predictive factors. Kaplan-Meier curve analysis revealed that patients with elevated SII levels experienced reduced 5-year DFS and OS rates compared to those with lower levels. In contrast, patients with elevated PNI levels demonstrated higher 5-year DFS and OS rates. Furthermore, a separate study highlighted that among EGC patients undergoing endoscopic submucosal dissection (ESD), those with higher PNI levels experienced significantly better OS rates than their counterparts with lower PNI levels.⁴⁶

In this study, the predictive capacity of SII was observed to surpass that of PNI and LNM. The combined use of these factors holds clinical relevance in prognosis prediction. These findings affirm the utility of PNI and SII in forecasting outcomes for EGC patients following radical gastric surgery. Moreover, PNI and SII could serve as valuable adjuncts to TNM staging, assessing gastric cancer patient survival, and guiding decisions regarding postoperative treatments.

In this study, subgroups of EGC patients were analyzed, revealing that elderly patients with lower PNI levels exhibited a poorer prognosis compared to those with higher PNI levels. Conversely, no statistically significant difference was observed between these groups in non-elderly patients. These findings suggest that PNI could serve as a potential prognostic indicator, specifically in elderly patients with EGC.

In addition, our study has several limitations. Firstly, it was conducted as a single-center retrospective analysis, which may introduce selection bias. Nevertheless, the substantial size of our study enhances the robustness of our evaluation regarding the prognostic significance of inflammation and nutrition indicators. Secondly, the absence of consensus on cutoff values for these indicators presents another challenge. In our investigation, we established optimal cutoff values for SII, NLR, PLR, ALI, and PNI using t-ROC curves. Therefore, future large-scale prospective studies are warranted to validate these findings comprehensively.

Conclusion

In conclusion, SII, PNI, and LNM emerge as robust predictors of survival outcomes following radical gastrectomy for EGC, demonstrating significant predictive efficacy.