| Background:In epidemiology, when evaluating the effect of the observed exposure on the outcome, if there were some variables are not only associated with the observed exposure, but also associated with the outcome variable, and these variables are not the intermediate variables on the chain of causation between the exposure and outcome, then these variables had caused confounding effects, we call these variables are confounders.The traditional approaches to handle these confounding effects, such as matching, stratification and regression adjust are often used. However, when there are several confounding factors while the sample size is relatively small, those traditional approaches are often inadequate. In recent years, the application of propensity score method makes the confounding effect more effectively handled.Immune and inflammatory factors in tumor microenvironment, are two key aspects in association with tumor prognosis. In current researches on immune and inflammatory cancer prognostic factors, investigators often focused on a specific immune factors or inflammatory factors to study their impact on prognosis, baseline characteristics considered are often limited to clinical tumor stage, the patient’s age, gender etc.Clinically, among a variety of immune or inflammatory factors are associated with each other, if we consider only a certain immune or inflammatory factor, the estimated effects would likely be biased.Currently, the outcome prediction in cancer is usually achieved by evaluating tissue samples obtained during surgical removal of the primary tumor focusing on their histopathological characteristics. Tumor staging (AJCC/UICC-TNM classification) summarizes data on tumor burden (T), presence of cancer cells in draining and regional lymph nodes (N), and evidence for metastases (M). However, this classification provides limited prognostic information in estimating the outcome in cancer and does not predict response to therapy.In the existing researches on the immune or inflammatory factors, researchers mostly base on a single immune or inflammatory factor to study its relation with prognosis. Only a few studies have considered a number of indicators to construct a prognostic score. We guess that prognostic score and tumor TNM stage are highly related, and also the tumor stage and prognosis, so these prognostic score probably also associated with tumor prognosis. The problem is, when in the same stage, the prognosis score is still related with the prognosis? In other words, is the prognosis score an independent prognostic factor with the prognosis of cancer patients?Objective:We hope to explore immune and inflammatory factors that affect the prognosis of gastric cancer patients, and using propensity score methods to correct some observed variables which caused baseline imbalance problem in observational studies. And try to build a multi-index weighted score by approximation method, in order to construct an immune or inflammatory factor score and consider the impact of the score on the prognosis on gastric cancer patients.Methods and Content:Data were retrospectively collected from patient electronic medical records, paper charts, pharmacy records from ZhongShan Hospital, and by contacting patients as needed. Electronic medical records, including liver and kidney function, tumor markers, CD cell subsets test results are retrospectively collected for each patient hospitalized every time. The patients’prognosis information is collected by telephone follow-up forms, including PFS and OS.We use the baseline data to explore the effect of baseline immune and inflammatory factors on the prognosis of gastric cancer. In this study, albumin, A/G and CD cell subsets, including CD19, CD3, CD8, CD4, etc. are considered immune markers, while white blood cell count, high-sensitivity C-reactive protein, NLR, PLR and mononuclear cell counts are considered as inflammatory markers. Survival curves were plotted by Kaplan-Meier method. Cox proportional hazards models were used to explore factors associated with prognosis. Using propensity score methods adjust the potential confounders.Using stochastic approximation method by R software, try to build a multi-index weighted score to evaluate the relation of the immune or inflammatory score with prognosis.Main results:1) After adjustment for gender, age and clinical stage of gastric cancer patients and some confounding factors, the impact of albumin on the prognosis of gastric cancer is still statistically significant, the HR= 0.87,95% CI (0.79,0.95), P=0.002. And when PFS was the endpoint, the result was HR= 0.97,95% CI= (0.90,1.05), P= 0.448, which was not statistically significant.2) Inflammatory indicators, such as PLR, NLR, CRP and monocyte counts were associated with the prognosis of gastric cancer, the higher the inflammatory index, the worse the prognosis of gastric cancer.3) Construction of inflammatory factors score is: Score=0.531*WBC+0.008*CRP+0.442*monocytes+0.010*NLR+0.009*PLRConclusion:After controlling for relevant confounding factors, immune and inflammatory factors, albumin, PLR, NLR, CRP and monocyte counts are associated with gastric cancer prognosis, and the higher the better prognosis of albumin, the higher inflammatory markers, the prognosis of gastric cancer the worse. This can improve the prognosis of gastric cancer patients in clinical practice to provide new non-target interventions. |
PDF Full Text Request