基于CT的放射组学特征预测肺腺癌EGFR突变状态

Prediction of EGFR mutation in lung adenocarcinoma based on CT radiomics features

目的:构建临床和放射学特征相结合的列线图以预测EGFR突变状态。方法:利用临床和放射学特征建立临床-放射学(C-R)模型。C-R模型与最佳放射组学模型建立临床-放射学-放射组学(C-R-R)模型。模型的预测性能使用ROC曲线进行评估。结果:在放射组学中10个特征与EGFR突变有关,使用支持向量机(SVM)分类器可获得最佳放射组学特征模型。C-R-R模型在预测EGFR突变状态方面具有最佳区分能力,其AUC在训练集和验证集中分别为0.95(95%CI 0.90~0.98)和0.94(95%CI 0.87~0.98)。结论:基于CT的放射组学特征与临床和放射学特征相结合而建立的非侵入性C-R-R模型可能为肺癌靶向治疗提供有价值的生物学信息。

Objective:To construct a nomogram based on a combination of radiomics features with clinical and image features to predict the EGFR mutation.Methods:The clinical features and radiological images features were used to establish a clinical-radiology(C-R)model.The C-R model with the best radiomics model were used to establish the clinical-radiology-radiomics(C-R-R)model.The predictive performance of models was evaluated by ROC curve.Results:The best radiomics model with 10 radimoics features was obtained by using the SVM classifier.The C-R-R model had the best distinguishing ability for predicting the EGFR mutation,with the AUC of 0.95(95%CI 0.90~0.98)in the training set and 0.94(95%CI 0.87~0.98)in the validation set.Conclusion:The noninvasive C-R-R model that combines clinical features,radiological image features and radiomics features can provide a useful image-based biological information for targeted therapy candidates in lung cancer.