基于PET/CT的非小细胞肺癌EGFR基因突变的预测

Predicting EGFR Mutation Status Based on PET/CT Imaging in Patients with Non-small Cell Lung Cancer

肺癌是癌症中极为高发且死亡率极高的一种。近年来,针对EGFR突变的靶向治疗方式取得了良好的临床效果,但目前的EGFR突变状态的检测方式主要是基于肿瘤样本的活检,不仅样本获取困难,还可能会带给病人极大的痛苦。因此提出了一种基于深度学习的非小细胞肺癌EGFR基因突变的预测方法,使用Resnet50作为主干网络建立深度学习模型,通过对图像特征、临床特征和影像组学特征的有效融合,实现对EGFR基因突变的预测。所提出的方法取得了0.831的AUC。

Lung cancer is one of the most common cancers with high incidence and mortality. In recent years, targeted therapy for EGFR mutations has achieved good clinical results. However, the current detection method of EGFR mutation status is mainly based on biopsy of tumor samples, which is not only difficult to obtain samples, but may also bring great pain to patients.Therefore, this paper proposes a deep learning-based method for EGFR mutation prediction in non-small cell lung cancer, using Resnet50 as the backbone network to establish a deep learning model to achieve the prediction of EGFR mutation through the effective fusion of image features, clinical features and radiomics features. The proposed method achieved an AUC of 0.831.