从TACE术前增强CT中，高通量地提取影像信息，进行感兴趣区域分割后，

从TACE术前增强CT中，高通量地提取影像信息，进行感兴趣区域分割后，获取肿瘤的影像组学特征总数。使用LASSO回归算法筛选有意义的组学特征，并建立放射组评分模型。根据临床评分和影像组学评分建立综合评分模型。使用ROC曲线下面积，比较三种模型（影像组学评分、临床评分和联合评分模型）的预测性能，并根据最优预测模型建立列线图。

From TACE preoperative enhanced CT, high-throughput image information is extracted, and the total number of imaging features of the tumor is obtained after segmentation of the region of interest. Lasso regression algorithm was used to screen meaningful omics features, and the radiation group scoring model was established. A comprehensive scoring model was established according to the clinical score and imaging score. The area under ROC curve was used to compare the prediction performance of three models (imaging omics score, clinical score and combined score model), and the nomogram was established according to the optimal prediction model.

High-throughput extraction of image information from preoperative enhanced CT of TACE, and segmentation of region of interest to obtain the total number of imaging omics features of tumor. LASSO regression algorithm is used to screen meaningful histologic features, and a radiation group scoring model is established. A comprehensive scoring model was established according to clinical score and imaging group score. Using the area under ROC curve, the prediction performance of three models (imaging group score, clinical score and joint score model) was compared, and nomogram was established according to the optimal prediction model.