Risk-stratification potential of inflammatory and nutritional biomarkers in gastric cancer: Insights from a propensity score-matched analysis.

Background/AimThis study aimed to evaluate the discriminatory performance of routinely accessible hematological, inflammatory, and nutritional biomarkers in differentiating patients with gastric cancer (GC) from healthy controls, and to examine their contribution within multivariable models after adjustment using Propensity Score Matching (PSM).MethodsA retrospective analysis was conducted using laboratory and demographic data from 109 GC patients and 152 healthy controls between 2020 and 2025. To reduce age- and sex-related imbalance between groups, 1:1 PSM was performed. Receiver Operating Characteristic (ROC) curve analysis was used to assess discriminatory performance, and exploratory multivariable logistic regression models were constructed to identify independent marker combinations associated with GC.ResultsAfter PSM, GC patients exhibited significantly lower hemoglobin, hematocrit, albumin, lymphocyte counts, and PNI values, alongside significantly higher platelet counts and inflammatory indices (SII, PLR, MLR, NLR, PIV) (all p<0.001). Among individual biomarkers, hemoglobin (AUC=0.830) and PLR (AUC=0.806) demonstrated the highest discriminatory performance. In exploratory multivariable models, reduced Hb, elevated PLT, low ALB, and low LYM were associated with GC in Model 1 (Nagelkerke R2=0.597; AUC=0.89), while PLR, PNI, and SII were associated with GC in Model 2 (Nagelkerke R2=0.492; AUC=0.839).ConclusionRoutinely available hematologic, inflammatory, and nutritional markers may serve as practical adjunctive tools for pre-endoscopic risk stratification in gastric cancer. Prospective validation in clinically representative populations is needed before broader clinical integration.