Background

Oxidative stress and immunoinflammation are two important areas in the field of cancer research. At present, no studies have confirmed that both can accurately predict the progression and long-term prognosis of gastric cancer (GC).

Methods

In this study, univariate and multivariate Cox proportional hazards regression analyses were used to analyze oxidative stress and immunoinflammatory indicators in the training set, and the best biomarkers for the establishment of an improved systemic prognosis score (mSPS) were screened. The Area Under the Curve (AUC) and Concordance Index (C-index) were used to evaluate the predictive ability of our model for long-term efficacy. Kaplan–Meier (K-M) survival analysis was conducted to assess survival outcomes. Univariate and multivariate Cox proportional hazards regression analyses were performed to investigate the impact of the model on tumor progression, lymph node metastasis, and overall prognosis. These findings were subsequently validated using both internal and external validation sets.

Results

A total of 3788 GC patients were included in this study, including 2339 in the training set, 1003 in the internal validation set, and 446 in the external validation set. The Receiver Operating Characteristic (ROC) curve analysis revealed that the AUC for mSPS in predicting overall survival (OS) among GC patients significantly surpassed that of all other markers related to oxidative stress and immunoinflammatory indicators (P < 0.001). Compared to the low mSPS group, patients in the high mSPS group were predominantly male and were more likely to be aged over 65, exhibit a higher BMI, possess a higher ASA grade, display a greater prevalence of vasculature and neuroinfiltration, exhibit deeper tumor invasion, have a higher incidence of lymph node metastasis, be diagnosed at a more advanced tumor stage (Stage III), and receive a higher proportion of chemotherapy (P < 0.05). Logistic regression analysis demonstrated that elevated mSPS was an independent risk factor for tumor invasion and lymph node metastasis in patients with GC. COX regression analysis further indicated that high mSPS independently contributed to a poorer prognosis among GC patients. The association between mSPS and pathological stage, as well as long-term prognosis, was validated through both internal and external sets, using the same cut-off value. The results aligned consistently with the training set effect.

Conclusions

As a newly developed prognostic score based on oxidative stress and immunoinflammatory indicators, mSPS can independently and effectively predict the tumor burden of advanced gastric cancer, and further distinguish the long-term prognosis of GC patients based on a pTNM staging system.