Predicting Response to Transarterial Chemoembolization in Hepatocellular Carcinoma Using Machine Learning Models
Niharika Dutta1, Pankaj Gupta1
1Department of Radiodiagnosis and Imaging, Postgraduate Institute of Medical Education and Research, Chandigarh, India.
The Indian Journal of Radiology & Imaging
|March 23, 2026
Summary
Machine learning models can predict transarterial chemoembolization (TACE) response in hepatocellular carcinoma (HCC). The clinical model showed the best performance, indicating potential for improved HCC treatment planning.
Area of Science:
- Oncology
- Radiology
- Artificial Intelligence
Background:
- Hepatocellular carcinoma (HCC) is a major cause of cancer mortality.
- Transarterial chemoembolization (TACE) is a primary treatment for intermediate-stage HCC.
- Predicting TACE response is crucial but challenging.
Purpose of the Study:
- To investigate machine learning (ML) models for predicting TACE response in HCC patients.
- To compare the performance of clinical, radiomic, deep neural network (DNN), and combined models.
Main Methods:
- Utilized the public WAW-TACE dataset with clinical data and CT images.
- Trained four models: clinical, radiomic, DNN, and combined clinicoradiological.
- Evaluated models using cross-validation and a held-out test set to predict TACE failure.
Main Results:
- The clinical support vector machine model achieved 70% accuracy and an AUC of 0.778.
- The radiomic logistic regression model showed 76.1% accuracy and an AUC of 0.740.
- The DNN and combined models had lower predictive performance in this cohort.
Conclusions:
- A multimodal approach was used to predict TACE response in HCC.
- Further optimization and multicenter data are needed to improve predictive accuracy.
Keywords:
computed tomographydeep neural networkhepatocellular carcinomamachine learningtransarterial chemoembolizationtreatment response
