Application of enhanced MRI-based radiomics nomogram in predicting the efficacy of initial TACE in patients with intermediate to advanced hepatocellular carcinoma
10.3969/j.issn.1008-794X.2025.10.008
- VernacularTitle:增强MRI的影像组学列线图预测中晚期肝细胞癌患者首次TACE的疗效
- Author:
Weiyue CHEN
1
;
Guihan LIN
;
Yongjun CHEN
;
Changsheng SHI
;
Jianfei TU
;
Jiansong JI
Author Information
1. 323000 浙江丽水 温州医科大学附属第五医院放射科全省影像与介入医学重点实验室
- Keywords:
hepatocellular carcinoma;
radiomics;
chemoembolization;
magnetic resonance imaging
- From:
Journal of Interventional Radiology
2025;34(10):1081-1088
- CountryChina
- Language:Chinese
-
Abstract:
Objective To discuss the application of enhanced MRI-based radiomics nomogram in predicting the efficacy of initial transcatheter arterial chemoembolization(TACE)in patients with intermediate to advanced hepatocellular carcinoma(HCC).Methods A total of 195 patients with advanced HCC(CNLC Ⅱ b-Ⅲb),who received initial TACE at the Affiliated Fifth Hospital of Wenzhou Medical University(Center 1)from January 2019 to March 2024,at the Lishui Municipal People's Hospital(Center 2)from July 2021 to June 2023,and at the Rui'an Municipal People's Hospital(Center 3)from January 2022 to January 2024,were enrolled in this study.A total of 134 patients from Center 1 were randomly divided into a training set(n=94)and an internal validation set(n=40)at a 7∶3 ratio;and other 61 patients from Center 2 and Center 3 were selected as the external validation set.Based on the modified Response Evaluation Criteria in Solid Tumors(mRECIST)criteria,the early efficacy of the initial TACE procedure was evaluated.The patients were divided into an effective group and an ineffective group.The tumor contours were delineated on the arterial,portal,and equilibrium phase images of enhanced MRI,and the corresponding radiomics features were extracted.Based on reduced-dimensional features,the Logistic regression,support vector machine,lightweight gradient boosting machine,and multi-layer perceptron models were established.Univariate analysis and multivariate logistic regression analysis were used to screen independent predictive factors,and a nomogram was established in conjunction with the optimal radiomics score.The area under the receiver operating characteristic curve(AUC)was used to evaluate the performance of the model,and decision curve analysis was adopted to calculate the net benefits.Results After screening,9 key radiomics features were obtained.The lightweight gradient boosting machine model showed good prediction performance.The AUCs of the training set,internal validation set,and external validation set were 0.909,0.836 and 0.783 respectively,which was selected as the optimal radiomics model.The nomogram constructed based on AFP level,peritumoral enhancement,and optimal radiomics score could further improve its performance,with AUC values of 0.962,0.890 and 0.821 in the training set,internal validation set,and external validation set respectively.Decision curve analysis showed that this model could bring higher net benefits to patients.Conclusion The nomogram constructed based on the enhanced MRI-based radiomics combined with AFP level and peritumoral enhancement can effectively predict the efficacy of the initial TACE in patients with intermediate to advanced HCC.
