Xiao Yafei, Li Peixin, Liang Siyuan
Objective To construct a nomogram model for predicting the 3-year tumor-specific mortality risk of patients with stage II-III gastric cancer after radical resection based on nutritional inflammation metabolism-related indicators, and to explore its potential clinical value, in order to provide evidence-based basis for individualized prognosis assessment and management of gastric cancer patients. Method A total of 363 patients with stage II-III gastric cancer who underwent radical resection at Beijing Friendship Hospital, Capital Medical University from January 2018 to December 2022 were retrospectively selected as the research subjects. They were randomly divided into a training set (254 cases) and a validation set (109 cases) at a ratio of 7:3. The training set was divided into a death group (91 cases) and a survival group (163 cases) based on the prognosis outcome. Clinical baseline data and nutritional, inflammatory, and metabolic-related laboratory indicators of the patients were collected. Cox proportional hazards regression analysis was used to screen the independent influencing factors of tumor-specific death in gastric cancer patients. Based on the screening results, a prognostic nomogram model was constructed. The discrimination, goodness-of-fit, and clinical application value of the model were verified through consistency index (C-index), time-dependent ROC curve, Gronnesby-Borgan test, Bootstrap self-repeated sampling, and decision curve analysis (DCA). Result Cox proportional hazards regression analysis showed that Eastern Cooperative Oncology Group performance status (ECOG-PS)of 2-3 (HR=2.074, 95%CI=1.351-3.186,P=0.001), neutrophil-to-lymphocyte ratio (NLR)≥2.66 (HR=1.753,95%CI=1.132-2.714,P=0.012), platelet-to-lymphocyte ratio (PLR) ≥ 151.72 (HR=2.258,95%CI=1.456-3.502,P<0.001), albumin-bilirubin (ALBI) grade 2 (HR=2.057, 95%CI=1.342-3.152,P=0.001), and prognostic immune-inflammation-nutrition index (PIIN) ≥ 31.49 (HR=2.120, 95%CI=1.366-3.291, P=0.001) were independent risk factors for tumor-specific death in gastric cancer patients; albumin/globulin ratio (AGR) ≥ 1.46 (HR=0.415,95%CI=0.263-0.654,P<0.001), and prognostic nutrition index (PNI) ≥ 45 (HR=0.571, 95%CI=0.357-0.912, P=0.019) were independent protective factors for tumor-specific death in gastric cancer patients. The C-index of the prognostic nomogram model based on the above indicators for predicting tumor-specific death in gastric cancer patients was 0.850 (95%CI=0.799-0.901); the AUC of 1-year, 2-year, and 3-year survival in the training set was 0.729, 0.783, and 0.829, respectively, and in the validation set were 0.590, 0.839, and 0.829; the Gronnesby-Borgan test showed that the fitting goodness of the training set (χ2=8.762,P=0.271) and the validation set (χ2=9.125,P=0.244) was good; the internal validation and decision curve of the Bootstrap method showed that the net benefit interval of the model for predicting tumor-specific death in gastric cancer patients was large, which was helpful for clinical decision-making. Conclusion The prognostic nomogram model constructed based on multiple-dimensional indicators of nutrition, inflammation, and metabolism in gastric cancer patients has good predictive efficacy and clinical practicability, can stratify the prognosis risk of gastric cancer patients, and provides scientific basis for the formulation of individualized intervention strategies such as perioperative nutritional support, anti-inflammation, and metabolic regulation for gastric cancer patients.
