Abstract: ObjectiveTo explore the application of multifactor dimensionality reduction(MDR) in the risk assessment of inflammatory cytokine gene (interleukin-1IL-1,tumor necrosis factor TNF and macrophage migration inhibitory factor MIF)-environment interaction in gastric cancer research. MethodsThe MDR software was applied to analyze gene-environment interaction on the risk of gastric cancer susceptibility,while the results of logistic regression was analysed as supplement. ResultsThe interaction among IL-1B-511、IL-1RN、TNF-A-308 genotypes and tumor familial history made the best MDR model with a statistical significance(P <0.01).Testing balance accuracy of this model was 0.72 and the cross-validation consistency was 10/10. The research subjects were divided into "high-risk" and "low-risk" groups according to this model.The estimated odds ratio (OR) value of interaction effects on gastric cancer was 13.49(95%confidence interval CI:8.62-21.10) with a statistical significance (P<0.01).There were significant differences among group 1(one factor),group 2(two factors) and group 3(three factors) (P<0.05) with the OR values of 2.17,11.61,and 27.19. ConclusionThe interaction of IL-1B-511、IL-1RN、TNF-A-308 and tumor familial history may increase the risk of gastric cancer.