Abstract: Objective The aim of this study was to analyze the predictive value of an immune scoring model based on immune cell infiltration for the efficacy and prognosis of immunotherapy in early gastric cancer. Methods The gene expression data and related clinical parameters of 167 early gastric cancer patients were downloaded from the Cancer Genome Atlas(TCGA)as the training set.The clinical information of 92 early gastric cancer patients who first visited our hospital from January 2017 to January 2020 was collected as a validation set.The infiltration of 22 immune cells in tumor tissues was calculated by the Cibersort software.The key predictive factors for early gastric cancer patients were further screened and determined by lasso analysis and multivariate Cox regression analysis.The eligible immune cells were used to construct the prognostic risk scoring model and verify it.Based on the risk model,a nomograph model was established to predict the probability of death and treatment failure in early gastric cancer patients,and evaluate the differentiation,accuracy,and reliability of the model. Results There was no statistically significant difference in the general information of patients between the training set and the validation set(P＞0.05).Compared with normal tissues,the content of CD8⁺T cells,activated memory CD4⁺T cells,M0 macrophages,M1 macrophages,resting dendritic cells,and activated dendritic cells in early gastric cancer tumor tissues decreased.The contents of regulatory T cells(Treg cells)and eosinophils increased,and the difference was statistically significant(P＜0.05).The LASSO analysis further screened five types of infiltrating immune cells,including activated memory CD4⁺T cells,M1 macrophages,Treg cells,activated dendritic cells and eosinophils.The survival curve results showed that the immune scoring model could effectively distinguish the survival time of patients.The total score of the prognostic nomogram model was 274,corresponding to risk of death of 72%.The total score of the treatment effect nomograph model was 307,corresponding to a 75% risk of ineffective treatment.The model validation results showed that the nomograph model has high discrimination,accuracy and reliability. Conclusion The immune scoring model based on immune cell infiltration has a certain predictive value for the efficacy and prognosis of immunotherapy in patients with early gastric cancer.

Key words: Immune cell infiltration, Immune scoring model, Early gastric cancer, Therapeutic effects, Prognosis