Immune-nutritional score-based prognostic model for predicting postoperative complications and survival in gastric cancer

Abstract

BACKGROUND

Postoperative complications are common in patients with gastric cancer (GC) and are closely associated with poor prognosis; however, effective tools for preoperative risk stratification remain limited.

AIM

To investigate the predictive value of the immune-nutritional score (INS) for postoperative complications and survival outcomes in patients with GC.

METHODS

We retrospectively reviewed the medical records of 247 patients with GC who underwent curative surgery between January 2019 and December 2022. Postoperative adverse events were categorized using the Clavien-Dindo grading system. Receiver operating characteristics (ROC) curve analyses were performed to evaluate the predictive value of the neutrophil-lymphocyte ratio (NLR), C-reactive protein (CRP), serum albumin (Alb), and lymphocyte count for postoperative complications. Based on cutoff values, 155 patients were assigned to the low-INS group and 92 to the high-INS group. Overall survival between groups was assessed using Kaplan-Meier survival analysis.

RESULTS

Significant differences were found in body mass index, lymphocyte count, neutrophil count, NLR, Alb, and CRP. Lymphocyte count and Alb were protective factors, whereas neutrophil count, NLR, and CRP were independent risk factors for complications. CRP and Alb showed relatively high predictive performance (area under the receiver operating characteristic curve = 0.9047 and 0.7854, respectively), while INS demonstrated the highest predictive value (area under the receiver operating characteristic curve = 0.9710, 95% confidence interval: 0.9515-0.9904). The incidence of postoperative complications was higher in the high-INS group, and overall survival was lower in the high-INS group (P < 0.001).

CONCLUSION

INS is an effective predictor of postoperative complications and survival outcomes in patients with GC, with strong clinical applicability.

Key Words: Gastric cancer; Immune-nutritional score; Postoperative complications; Prognosis; Survival outcomes; Inflammatory markers; Risk prediction

Core Tip: This study proposes an immune-nutritional score, derived from routinely available laboratory parameters, to evaluate the risk of postoperative complications and survival in patients with gastric cancer. By integrating indicators of nutritional status, immune function, and systemic inflammation, immune-nutritional score showed improved predictive performance compared with single markers. It may provide a practical approach for preoperative risk assessment and support individualized perioperative management. Further validation in larger, prospective cohorts is warranted to confirm its generalizability and clinical utility.

INTRODUCTION

Globally, cancer remains a major public health challenge and is a leading cause of morbidity and mortality. Among malignancies, gastric cancer (GC) is highly prevalent and particularly burdensome in East Asia[1]. Although surgical techniques and perioperative management have advanced in recent years, radical surgery remains the primary treatment for GC. However, the incidence of postoperative complications remains high and directly affects long-term survival outcomes and quality of life[2]. Therefore, effectively identifying high-risk patients preoperatively and implementing individualized perioperative management are critical clinical priorities.

In recent years, the immunonutritional score has been proposed as a composite indicator reflecting the interplay among systemic inflammation, immune status, and nutritional condition. It typically includes parameters, including serum albumin (Alb), neutrophil-to-lymphocyte ratio (NLR), and C-reactive protein (CRP), which are routinely measured and widely used to assess prognosis in various malignancies. Host nutritional and immune status plays a critical role in postoperative recovery and prognosis in patients with GC. Alb, NLR, and CRP are associated with postoperative complications and survival outcomes[3]. However, individual parameters have limited predictive performance and do not comprehensively reflect immune-inflammatory responses and nutritional status[4].

The interaction between systemic inflammation and immune response plays a key role in tumor progression and postoperative outcomes in GC. Dysregulated inflammation may impair immune function, delay wound healing, and increase susceptibility to complications, thereby adversely affecting both short-term recovery and long-term prognosis. Recently, prediction models integrating multiple parameters have gained increasing attention. Although some scoring systems have been preliminarily applied in gastrointestinal malignancies, a unified immune-nutritional assessment tool with high sensitivity and specificity remains lacking[5,6].

Against this background, this study develops the INS, integrating multiple immune-inflammatory and nutritional indicators to improve prediction of postoperative complications and survival outcomes in patients with GC. Unlike single indicators or localized models, INS was establishing by determining optimal cutoff values through receiver operating characteristic (ROC) curve analysis, followed by patient stratification and validation using Kaplan-Meier survival analysis. This approach may facilitate risk stratification and individualized intervention while providing evidence to optimize perioperative management and improve long-term outcomes. Therefore, this study aimed to evaluate the predictive value of INS for postoperative complications and survival outcomes in patients with GC and to assess its clinical utility.

MATERIALS AND METHODS

Study population

This study adopted a retrospective cohort design, using patient data from the Affiliated Hospital of Xuzhou Medical University GC surgery database. A total of 247 consecutive patients who met the predefined eligibility criteria and underwent radical surgery for GC between January 2019 and December 2022 were included. Referring to the study by Mao et al[7], the sample size was of a comparable order of magnitude, meeting statistical power requirements. Patients were divided into a complication cohort (n = 82) and a non-complication cohort (n = 165) based on the occurrence of postoperative complications. Complication severity was graded using the Clavien-Dindo classification, with grade ≥ I defined as the presence of complications. The Ethics Committee of the Affiliated Hospital of Xuzhou Medical University approved this study protocol. All data were anonymized before analysis, and the study strictly followed the Declaration of Helsinki.

Inclusion criteria: (1) Individuals meeting the diagnostic criteria for GC according to the Guidelines for the Diagnosis and Treatment of GC and undergoing radical surgery[8]; (2) Age ≥ 18 years; (3) Complete clinical data; and (4) Complete follow-up data with available survival outcomes.

Exclusion criteria: (1) Patients with concurrent malignancies or unknown primary tumor site; (2) Patients with severe hepatic or renal dysfunction or immune system diseases; and (3) Patients with clinically evident acute infection or active inflammatory disease.

Study methods

Data collection: Data were extracted from electronic health records and the follow-up database, including: (1) Baseline characteristics [e.g., sex, age, body mass index (BMI), smoking and alcohol history, and comorbidities]; (2) Clinicopathological data [tumor location and size, histological differentiation, tumor-node-metastasis (TNM) stage, lymph node metastasis, perineural invasion, and vascular invasion]; (3) Laboratory parameters [preoperative serum Alb, CRP, lymphocyte count, neutrophil count (NEU), and NLR, calculated as NEU/Lymphocyte count]; and (4) Outcome measures [postoperative complications (types and grades) and survival status].

Classification of postoperative complications: Postoperative events within 30 days of surgery were recorded, including bleeding, anastomotic leakage, bowel obstruction, surgical site infection, pulmonary infection, and other complications. Complication severity was graded from I to V according to the Clavien-Dindo classification[9]. In this study, grade ≥ I was defined as a complication. For patients with multiple complications, the highest grade was recorded.

INS construction: The INS was constructed based on the results of multivariate logistic regression analysis. Before model development, continuous variables (NEU, CRP, Alb, and lymphocyte count) were standardized using z-score transformation to reduce scale differences and improve comparability. The z-score was calculated as follows: Z = (X - μ)/σ, where μ and σ represent the mean and standard deviation of each variable derived from the present study cohort. Variables identified as independent predictors of postoperative complications were included in the final model. The INS was calculated as a weighted sum of these variables, with weights derived from standardized regression coefficients (standardized β values), thereby reflecting the relative contribution of each predictor. The final model was expressed as a logistic regression equation, and the corresponding mean and standard deviation used for standardisation are provided in Table 1 to ensure reproducibility. ROC curve analysis was then performed to assess the predictive performance, and the optimal cutoff value was determined using the Youden index. Based on this cutoff, patients were stratified into a low-INS group (n = 155) and a high-INS group (n = 92). The incidence of postoperative complications was compared between both groups.

Table 1 Mean and standard deviation of variables used for z-score standardisation in the immune-nutritional score model.

Variable	mean	SD
Lymphocyte	1.39	0.30
NEU	4.40	1.44
Alb	3.94	0.48
CRP	10.10	3.12

NEU: Neutrophil-to-lymphocyte ratio; Alb: Albumin; CRP: C-reactive protein.

Survival outcomes: OS was defined as the time from diagnosis to death from any cause or the end of follow-up. Kaplan-Meier survival curves were generated, and differences in OS between the low-INS and high-INS groups were evaluated using the log-rank test. The follow-up deadline was April 30, 2025.

Statistical analysis

Statistical analyses were performed using SPSS software (version 26.0; IBM Corp., Armonk, NY, United States) and GraphPad Prism (version 9.0; GraphPad Software, San Diego, CA, United States). The Shapiro-Wilk test was used to assess normality of continuous variables. Normally distributed variables are expressed as mean ± SD and compared using the independent samples t-test, whereas non-normally distributed variables are expressed as median (P₂₅, P₇₅) and compared using the rank-sum test. Categorical variables are expressed as n (%) and compared using the χ² test. Model predictive performance was evaluated using ROC curve analysis, with optimal cutoff values determined using the Youden index. Kaplan-Meier analysis was used for survival evaluation. A two-tailed P < 0.05 was considered statistically significant.

RESULTS

Baseline characteristics

No significant differences were observed between the complication group and the non-complication group in age, sex, smoking history, alcohol consumption, combined diabetes, maximum tumor diameter, tumor location, histological differentiation, TNM stage, gastrectomy type, lymph node metastasis, perineural invasion, vascular invasion, carcinoembryonic antigen, or alpha-fetoprotein, white blood cell count, or platelet count (all P > 0.05). However, BMI, lymphocyte count, NEU, NLR, serum Alb, and CRP differed significantly between groups (all P < 0.05; Table 2).

Table 2 Baseline characteristics of patients with gastric cancer, n (%)/mean ± SD.

Variables		Case	Non-complication group (n = 165)	Complication group (n = 82)	χ2/t/z	P value
Age	< 65 years	150	97 (58.79)	53 (64.63)	0.79	0.38
	≥ 65 years	97	68 (41.21)	29 (35.37)
Sex	Male	186	128 (77.58)	58 (70.73)	1.38	0.24
	Female	61	37 (22.42)	24 (29.27)
BMI	> 24 kg/m2	107	79 (47.88)	28 (34.15)	4.21	0.04
	≤ 24 kg/m2	140	86 (52.12)	54 (65.85)
Smoking history	Yes	106	68 (41.21)	38 (46.34)	0.59	0.44
	No	141	97 (58.79)	44 (53.66)
Alcohol consumption history	Yes	128	84 (50.91)	44 (53.66)	0.17	0.68
	No	119	81 (49.09)	38 (46.34)
Combined diabetes	Yes	87	53 (32.12)	34 (41.46)	2.10	0.15
	No	160	112 (67.88)	48 (58.54)
Neoadjuvant chemotherapy	Yes	46	28 (16.97)	18 (21.95)	0.90	0.34
	No	201	137 (83.03)	64 (78.05)
Maximum tumor diameter	> 5 cm	125	77 (46.67)	48 (58.54)	3.10	0.08
	≤ 5 cm	122	88 (53.33)	34 (41.46)
Tumor location	Gastric body	28	20 (12.12)	8 (9.76)	0.31	0.86
	Fundus and cardia	54	36 (21.82)	18 (21.95)
	Gastric antrum	165	109 (66.06)	56 (68.29)
Histological differentiation	Poorly differentiated	132	82 (49.7)	50 (60.98)	2.81	0.25
	Moderately differentiated	71	51 (30.91)	20 (24.39)
	Well differentiated	44	32 (19.39)	12 (14.63)
TNM stage	I-II	91	65 (39.39)	26 (31.71)	1.39	0.24
	III-IV	156	100 (60.61)	56 (68.29)
Type of gastrectomy	Partial	94	66 (40.0)	28 (34.15)	0.80	0.37
	Complete	153	99 (60.0)	54 (65.85)
Lymph node metastasis	Yes	107	65 (39.39)	42 (51.22)	3.12	0.08
	No	140	100 (60.61)	40 (48.78)
Perineural invasion	Yes	52	34 (20.61)	24 (29.27)	2.29	0.13
	No	195	131 (79.39)	58 (70.73)
Vascular invasion	Yes	102	64 (38.79)	38 (46.34)	1.29	0.26
	No	145	101 (61.21)	44 (53.66)
CEA	< 5 ng/mL	158	108 (65.45)	50 (60.98)	0.48	0.49
	≥ 5 ng/mL	89	57 (34.55)	32 (39.02)
AFP	≤ 20 ng/mL	219	145 (89.51)	74 (87.06)	0.33	0.56
	> 20 ng/mL	28	17 (10.49)	11 (12.94)
WBC (× 109/L)		-	6.85 ± 1.66	7.21 ± 1.78	1.59	0.11
PLT (× 109/L)		-	132.18 ± 9.82	133.90 ± 11.13	1.24	0.22
Lymphocyte (× 109/L)		-	1.44 ± 0.29	1.27 ± 0.29	4.24	< 0.001
NEU (× 109/L)		-	3.91 ± 1.14	5.39 ± 1.47	8.70	< 0.001
NLR, median (P25, P75)		-	2.73 (2.07, 3.45)	4.21 (3.38, 5.31)	7.77	< 0.001
Alb (g/dL)		-	4.09 ± 0.48	3.64 ± 0.32	7.61	< 0.001
CRP (mg/L)		-	8.59 ± 1.92	13.13 ± 2.88	14.70	< 0.001

BMI: Body mass index; TNM: Tumor-node-metastasis; CEA: Carcinoembryonic antigen; AFP: Alpha-fetoprotein; WBC: White blood cell count; PLT: Platelet count; NEU: Neutrophil count; NLR: Neutrophil-to-lymphocyte ratio; Alb: Albumin; CRP: C-reactive protein.

Multivariate analysis of risk factors for postoperative complications in patients with GC

Multivariate logistic regression identified lymphocyte count, NEU, NLR, Alb, and CRP as independent factors associated with postoperative complications in patients with GC. As NLR is derived from neutrophil and lymphocyte counts, it was excluded from the multivariate model to avoid multicollinearity. Specifically, lymphocyte count [odds ratio (OR) = 0.294, 95% confidence interval (CI): 0.155-0.557, P < 0.001] and Alb (OR = 0.249, 95%CI: 0.128-0.486, P < 0.001) were protective factors, whereas NEU (OR = 3.366, 95%CI: 1.801-6.293, P < 0.001) and CRP (OR = 24.238, 95%CI: 8.569-68.557, P < 0.001) were independent risk factors. BMI was not significantly associated with postoperative complications (OR = 0.787, 95%CI: 0.274-2.264, P = 0.658; Table 3).

Table 3 Multivariate analysis of risk factors for postoperative complications in patients with gastric cancer.

Factor	β	SE	Wald χ2 value	P value	OR	95%CI
Constant	-1.603	0.393	16.611	0	0.201	-
BMI	-0.239	0.539	0.197	0.658	0.787	0.274-2.264
Lymphocyte	-1.224	0.326	14.098	0	0.294	0.155-0.557
NEU	1.214	0.319	14.458	0	3.366	1.801-6.293
A1b	-1.39	0.341	16.584	0	0.249	0.128-0.486
CRP	3.188	0.53	36.114	0	24.238	8.569-68.557

OR: Odds ratio; CI: Confidence interval; SE: Standard error; BMI: Body mass index; NLR: Neutrophil-to-lymphocyte ratio; Alb: Albumin; CRP: C-reactive protein.

INS model construction and ROC analysis

Based on the multivariate logistic regression analysis, the final INS model was formulated as follows: Logit(P) = -1.603 - 1.224 × Z(Lymphocyte) + 1.214 × Z(NEU) - 1.390 × Z(Alb) + 3.188 × Z(CRP). Z was calculated as Z = (X - μ)/σ, where μ and σ represent the mean and standard deviation of each variable derived from the present study cohort (Table 3). The predicted probability was calculated as P = 1/{1 + exp[-logit(P)]}. ROC curve analysis showed that NEU, CRP, Alb, and lymphocyte count had predictive value for postoperative complications in patients with GC, with AUCs of 0.7891, 0.9047, 0.7854, and 0.6457, respectively (all P < 0.05). Among these, CRP exhibited the highest predictive performance (AUC = 0.9047, 95%CI: 0.8618-0.9476), followed by Alb (AUC = 0.7854, 95%CI: 0.7301-0.8407). Based on the results of the multivariate logistic regression with standardized variables, the INS was constructed using standardized regression coefficients (standardized β values) as weights. ROC analysis showed that INS had significantly superior predictive performance compared with individual indicators, with an AUC of 0.9710 (95%CI: 0.9515-0.9904), sensitivity of 93.90%, and specificity of 90.91%. The optimal cutoff value for the INS was determined as > -0.7155 using the Youden index, and patients were accordingly stratified into a low-INS group (n = 155) and a high-INS group (n = 92) (Table 4, Figure 1).

Figure 1 Receiver operating characteristics analysis on the predictive performance of immune-nutritional score. NLR: Neutrophil-to-lymphocyte ratio; Alb: Albumin; CRP: C-reactive protein; INS: Immune-nutritional score.

Table 4 Diagnostic performance of immune-nutritional score based on receiver operating characteristics analysis.

Variables	AUC	Sensitivity (%)	Specificity (%)	Optimal cutoff	95%CI
NEU	0.7891	67.07	83.08	> 4.85	0.7247-0.8535
Alb	0.7854	85.37	63.03	< 3.95	0.7301-0.8407
Lymphocyte	0.6457	46.34	73.94	< 1.25	0.5739-0.7174
CRP	0.9047	81.71	86.67	> 10.68	0.8618-0.9476
INS score	0.9710	93.90	90.91	> -0.7155	0.9515-0.9904

AUC: Area under the receiver operating characteristic curve; CI: Confidence interval; NEU: Neutrophil count; Alb: Serum albumin; CRP: C-reactive protein; INS: Immune-nutritional score.

Relationship between INS and postoperative complications in patients with GC

The incidence of postoperative complications was higher in the high-INS group than that in the low-INS group [83.70% (77/92) vs 3.23% (5/155); χ² = 168.57; P < 0.001)]. The high-INS group also had higher rates of Clavien-Dindo grade I-II (63.04% vs 2.58%) and grade ≥ III complications (20.65% vs 0.65%) (Table 5).

Table 5 Relationship between immune-nutritional score and postoperative complications, n (%).

Variables	Low-INS group (n = 155)	High-INS group (n = 92)	χ2	P value
Clavien-Dindo grade
Grade I-II	4 (2.58)	58 (63.04)	-	-
Grade ≥ III	1 (0.65)	19 (20.65)	-	-
Types of complications
Pulmonary infection	1 (0.65)	21 (22.83)	-	-
Surgical site infection	1 (0.65)	10 (10.87)	-	-
Anastomotic leakage	1 (0.65)	11 (11.96)	-	-
Postoperative hemorrhage	1 (0.65)	9 (9.78)	-	-
Intestinal obstruction	1 (0.65)	11 (11.96)	-	-
Intra-abdominal infection	0 (0.00)	7 (7.61)	-	-
Other complications	0 (0.00)	8 (8.7)	-	-
Incidence of postoperative complications	5 (3.23)	77 (83.7)	168.57	< 0.001

INS: Immune-nutritional score.

Relationship between INS and survival outcomes in patients with GC

Kaplan-Meier survival analysis showed that patients with high INS had significantly poorer survival rates than those with low INS (log-rank χ² = 30.70, P < 0.0001). The median follow-up time for the entire cohort was 33.2 months, and the median survival was 72 months in the low-INS group and 41 months in the high-INS group. These findings indicate that high INS is significantly associated with an unfavorable clinical prognosis (Figure 2).

Figure 2 Kaplan-Meier survival analysis of gastric cancer patients stratified by immune-nutritional score. INS: Immune-nutritional score.

DISCUSSION

Based on multivariate logistic regression and ROC curve analyses, we constructed an INS model. Overall, our findings indicate that low serum Alb and lymphocyte levels reflect impaired nutritional and immune function, whereas elevated CRP, NEU, and NLR reflect activation of systemic inflammation and stress responses; these factors significantly influence the occurrence of postoperative complications[10,11]. Consistently, baseline comparisons showed decreased levels of protective indicators (Alb and lymphocyte count) and increased levels of risk-related inflammatory markers (NEU, NLR, and CRP) in patients with postoperative complications, supporting their roles in perioperative risk stratification. By integrating these indicators, the INS model demonstrated superior predictive performance for both complications and OS compared with single indicators, suggesting its potential application in preoperative risk stratification and clinical intervention[6].

In recent years, nutritional and immune status have been widely recognized as key determinants of perioperative prognosis in GC[10]. Several studies have shown that malnutrition and systemic inflammation increase the risk of surgery-related complications and affect long-term survival by impairing immune function and tissue repair capacity[11]. Alb is a classical nutritional marker, with decreased levels reflecting impaired protein synthesis and chronic inflammatory burden. Mechanistically, hypoalbuminemia may indicate a catabolic state driven by tumor-associated inflammation, compromising tissue repair and increasing susceptibility to postoperative complications. Previous studies have shown that low Alb is closely associated with postoperative infections, anastomotic leakage, and prolonged hospitalization[12]. Lymphocyte count reflects immune responsiveness, with low levels indicating immunosuppression and reduced anti-infective capacity, thereby increasing complication risk[13]. Reduced lymphocyte levels may also impair anti-tumor immunity, leading to inadequate control of residual tumor cells and poorer long-term outcomes. Conversely, NEU and NLR indicate the interplay between inflammatory responses and immune suppression. Elevated NLR has been consistently linked to worse outcomes across multiple cancer types, including gastrointestinal tumors such as GC[14,15]. Increased NEU may promote tumor progression through the release of proinflammatory mediators and suppression of cytotoxic immune responses, contributing to both postoperative complications and tumor progression. CRP, as an acute-phase protein, directly reflects inflammatory load. In this study, it showed the highest predictive performance among individual indicators (AUC = 0.9047), highlighting its sensitivity in reflecting postoperative inflammatory burden and its strong association with postoperative complications[16]. Notably, CRP exhibited a relatively large effect size in the multivariate model (OR = 24.238), which may be related to its high sensitivity to systemic inflammatory burden and the modeling of continuous variables; however, this finding should be interpreted with caution and warrants further validation.

These findings indicate that INS can effectively capture the preoperative physiological status of patients with GC, thereby influencing postoperative recovery and long-term outcomes[17]. Notably, the INS model integrating these indicators demonstrated higher predictive performance compared with individual indicators (AUC = 0.9710). However, this high performance may be partly attributable to model development within a single cohort; therefore, external validation is warranted to confirm its generalizability. This finding aligns with recent trends in which multidimensional scoring systems increasingly replace single-parameter approaches[6], reflecting that the risk of complex diseases often arises from interactions among multiple factors, whereas single indicators cannot fully represent overall physiological status[18]. Our results show that INS is associated with both short-term complications and long-term survival outcomes; however, the prognostic value for survival should be further validated in cohorts with longer follow-up durations. This “dual predictive” advantage underscores the clinical significance of INS as a comprehensive perioperative risk assessment tool[19].

Potential mechanisms linking immune-nutritional status with postoperative complications and survival outcomes may involve multiple pathways. First, malnutrition may impair immune cell function, delay wound healing, and reduce anti-infective capacity, thereby increasing complication risk[20]. Second, persistent systemic inflammation may trigger immune imbalance and tissue damage through the release of proinflammatory cytokines such as interleukin-6 and tumor necrosis factor-alpha, exacerbating postoperative stress responses[21]. Third, immune and nutritional imbalance may influence the tumor microenvironment, promoting recurrence and metastasis and shortening long-term survival[22]. These mechanisms are consistent with the patterns observed in this study, in which indicators of nutritional and immune competence were reduced and inflammatory markers elevated in patients with postoperative complications, supporting the biological plausibility of INS.

From a clinical perspective, the INS model appears feasible and potentially generalizable. All required laboratory parameters are routinely measured, making it cost-effective and accessible across healthcare settings. INS stratification may help identify high-risk patients preoperatively and guide nutritional intervention, immune modulation, and individualized perioperative management[22]. By integrating immune, inflammatory, and nutritional dimensions, INS may provide a more comprehensive assessment of perioperative risk than single indicators. For instance, patients with high INS may benefit from enhanced nutritional support, inflammation control, and personalized perioperative care to reduce complication risk and improve long-term outcomes[23]. Additionally, INS may serve as a supplementary tool for postoperative follow-up and long-term prognostic evaluation, enabling more targeted risk assessment and intervention.

However, several limitations should be acknowledged. First, as a retrospective single-center study, there is potential for selection bias, and generalizability requires confirmation in multicenter prospective studies. Second, although INS integrates multiple indicators, other relevant factors, such as nutritional risk screening tools (e.g., NRS2002), body composition measures, and more detailed inflammatory biomarkers, were not included, limiting model comprehensiveness. Moreover, postoperative complications were analyzed as a composite outcome rather than stratified (e.g., infectious vs non-infectious), which may limit more refined evaluation of INS predictive performance. Third, the follow-up duration was limited and may be insufficient to fully assess very long-term survival. Additionally, INS performance across subgroups (e.g., TNM stages, surgical methods, and age groups) was not evaluated, which may limit the assessment of its generalizability in diverse clinical settings. Finally, INS should be applied alongside clinical scoring systems and individual patient factors to avoid oversimplification of complex clinical scenarios.

In conclusion, this study is the first to systematically validate the dual predictive value of INS for postoperative complications and long-term survival in patients with GC. By integrating routine clinical laboratory indicators, INS demonstrated superior predictive ability compared with single parameters, with good accessibility and clinical utility. It provides a novel tool for preoperative risk stratification, perioperative management, and individualized treatment. Future studies should validate its generalizability in multicenter cohorts and further assess its predictive performance across different complication types and patient subgroups to enhance clinical applicability.

CONCLUSION

The findings suggest that INS is a valuable tool for predicting postoperative complications and survival outcomes in patients with GC, with superior performance compared with individual markers. As a simple tool based on routine assessments, INS may provide valuable reference for preoperative risk stratification and personalized management.