基于CT影像组学预测肝细胞癌TACE术后疗效

Radiomics Analysis Based on CT for Prediction of the Therapeutic Effect to Transarterial Chemoembolization in Hepatocellular Carcinoma

目的构建临床-影像组学联合模型预测肝细胞癌(HCC)患者经导管肝动脉化疗栓塞(TACE)术后生存期。方法回顾性搜集经TACE治疗的HCC患者151例。在动、门静脉图像上构建肿瘤和瘤周的影像组学模型并分组组合。联合临床独立危险因素与影像组学模型构建临床-影像组学联合模型。利用C指数评估模型区分度,校准曲线评估模型校准度。结果动脉期瘤周3 mm与静脉期肿瘤的组合模型(A3+P)的预测效能与稳定性较好(C指数:训练集0.739、验证集0.732)。综合临床独立危险因素后联合模型预测性能有所提高(C指数:训练集0.773,验证集0.751)。结论临床-影像组学联合模型对预测TACE术后生存期具有价值。

Objective To develop and validate a clinical-radiomics model for predicting overall survival after transarterial chemoembolization(TACE)in patients with hepatocellular carcinoma(HCC).Methods This retrospective study included 151 patients with HCC who underwent TACE.The tumor and peritumor radiomic signatures were constructed on the arterial phase and portal venous phase images,and the signature were subsequently combined and evaluated for signature performance using the C-index.A combined clinical-radiomics model was constructed COX proportional hazards regression models,which combining clinical independent risk factors with radiomics signatures.Model discrimination was assessed using the C-index,and calibration was assessed using calibration plots.Results Out of the 151 patients enrolled in the study,112 died(74%)and 39 survived(26%).The best prediction performance was obtained for the combined radiomic signatures of arterial and venous phases(A3+P)(C-index:training set 0.739,validation set 0.732).The combined model after combining clinically independent risk factors and radiomic signature not only improved predictive performance(C-index:training set 0.773,validation set 0.751),but also classified patients into low-risk and high-risk groups with significant differences in survival.Conclusion The combined clinical-radiomics model was valuable in predicting survival after TACE.