Objective To evaluate the value of dual-energy CT combined with carcinoembryonic antigen (CEA)and soluble cytokeratin 19 fragment(CY21-1) in predicting epidermal growth factor receptor(EGFR) mutations of patients with advanced lung adenocarcinoma.Methods General informations,quantitative parameters of dual-energy computed tomography (DECT),CEA,CY21-1 and EGFR datas were collected of 90 patients between October 2019 and April 2022,the quantitative parameters include iodine concentration(IC),normalized iodine concentration(NIC) and slope of the curve(k) in venous phase. According to whether there was EGFR gene mutation or not,it was divided into mutation group and wild group.To compare the difference of clinical data and DECT parameters,logistic regression model was established to in the two groups Evaluation of model prediction efficiency by receiver operating characteristic(ROC) curve.Results Among all cases,55 cases were EGFR gene mutations mostly in female and non-smokers.IC,NIC,k in venous phase(VP) and CEA in mutation group were higher than another group,but CY21-1 is opposite.The area under ROC curve of IC(VP),k(VP),CEA and CY21-1 were 0.703,0.642,0.725,0.609,respectively.DECT parameters combinded to establish regression model 1,and then combinded CEA and CY21-1 to establish model 2, the area under curve(AUC) is 0.807, Model 2 was higher to model 1 and other single factors.There was no significant difference in age and NIC(VP)(P>0.05).Conclusion EGFR gene mutations tend to occur in women and non-smokers, IC,k in venous phase and CEA and CY21-1 can effectively predict EGFR gene mutation to some extent.