Objective To explore the application value of dual-energy computed tomography(DECT)of the chest combined with enhanced MRI of the brain in predicting anaplastic lymphoma kinase(ALK)gene mutation in patients with brain metastasis of lung adenocarcinoma.Methods The clinical and imaging data of 77 patients with brain metastasis of lung adenocarcinoma were analyzed retrospectively.Clinical data and DECT parameters of chest,enhanced MRI images of brain and ALK gene mutation status were included.And the quantitative parameters of DECT included normalized iodine concentration in arterial phase(NICA),normalized iodine concentration in venous phase(NICV),λ in arterial phase(λA)and λ in venous phase(λV).According to the ALK gene detection results,they were divided into ALK positive group(n=17)and ALK negative group(n=60).The differences of clinical data,DECT parameters,and MRI images between the two groups were compared,and the logistic regression model for predicting ALK gene mutation was constructed.The prediction of the model was evaluated by receiver operating characteristic(ROC)curve and area under the curve(AUC).Results Univariate analysis revealed that ALK positive were found in youngers and non-smokers.λV and NICV in ALK positive group were significantly lower than those in ALK negative group.The peritumoral brain edema index(PBEI)and peritumoral brain edema size(PBES)of ALK positive group were significantly smaller than those of ALK negative group.There was statistically significant in age,smoking history,λV,NICV,PBEI and PBES between the two groups(P＜0.05).Clinical features combined with DECT parameters were used to establish model 1,clinical features combined with DECT parameters and MRI parameters were used to establish model 2.Model 2 was significantly superior to model 1 or univariate analysis,respectively,in predicting ALK gene mutation by DeLong test,the AUC was 0.927.Conclusion ALK mutations are more common in youngers and non-smokers.λV,NICV,PBEI,and PBES are reliable predictors of ALK mutations.Model 2 can improve predictive efficiency significantly.

Objective To explore the application value of dual-energy computed tomography(DECT)of the chest combined with enhanced MRI of the brain in predicting anaplastic lymphoma kinase(ALK)gene mutation in patients with brain metastasis of lung adenocarcinoma.Methods The clinical and imaging data of 77 patients with brain metastasis of lung adenocarcinoma were analyzed retrospectively.Clinical data and DECT parameters of chest,enhanced MRI images of brain and ALK gene mutation status were included.And the quantitative parameters of DECT included normalized iodine concentration in arterial phase(NICA),normalized iodine concentration in venous phase(NICV),λ in arterial phase(λA)and λ in venous phase(λV).According to the ALK gene detection results,they were divided into ALK positive group(n=17)and ALK negative group(n=60).The differences of clinical data,DECT parameters,and MRI images between the two groups were compared,and the logistic regression model for predicting ALK gene mutation was constructed.The prediction of the model was evaluated by receiver operating characteristic(ROC)curve and area under the curve(AUC).Results Univariate analysis revealed that ALK positive were found in youngers and non-smokers.λV and NICV in ALK positive group were significantly lower than those in ALK negative group.The peritumoral brain edema index(PBEI)and peritumoral brain edema size(PBES)of ALK positive group were significantly smaller than those of ALK negative group.There was statistically significant in age,smoking history,λV,NICV,PBEI and PBES between the two groups(P＜0.05).Clinical features combined with DECT parameters were used to establish model 1,clinical features combined with DECT parameters and MRI parameters were used to establish model 2.Model 2 was significantly superior to model 1 or univariate analysis,respectively,in predicting ALK gene mutation by DeLong test,the AUC was 0.927.Conclusion ALK mutations are more common in youngers and non-smokers.λV,NICV,PBEI,and PBES are reliable predictors of ALK mutations.Model 2 can improve predictive efficiency significantly.