基于脑MRI的机器学习预测非小细胞肺癌T790M突变OA北大核心CSTPCD

Objective:In this study,an artificial intelligence model was established based on contrast-enhanced T1-weighted(T1C)and T2-weighted(T2W)sequences of brain MRI to predict drug-resistant T790M mutations in lung cancer brain metastasis patients undergoing targeted therapy.Methods:In this study,T1C and T2W MRI imaging data and clinical data of 80 lung cancer brain metastasis patients(from June 2017 to December 2019)were collected for retrospective analysis(the data was divided into training and validation cohorts in a ratio of 2∶1).The unsupervised k-means algorithm was used to segment the tumor region into high-brightness and low-brightness subregions,and the radiomics features of every subregion were extracted to establish a model to evaluate the diagnostic performance of every model.Receiver operating characteristic(ROC)curves were plotted,and the area under the curve(AUC),specificity and sensitivity were used as evaluation metrics to analyze the potential clinical application value of the model.Results:Statistical calculations combining T1C and T2W MRI and clinical features showed that the model established in this study had good predictive ability for T790M mutation,with AUCs of 0.899 and 0.818 in the training and testing sets,respectively.Conclusion:The computer model established in this study can effectively predict the T790M mutation in lung cancer brain metastasis patients and has potential clinical auxiliary diagnostic value.