Home›Oncology› Deep Learning MRI Models Predict Treatment Response in Hepatocellular Carcinoma Following Transarterial Chemoembolization
Deep Learning MRI Models Predict Treatment Response in Hepatocellular Carcinoma Following Transarterial ChemoembolizationAI Models Predict Liver Cancer Treatment Response Better Than Standard Scores
Frontiers in MedicinePublished April 28, 2026DOI ↗Editorial oversight: Dr. Julia Lee, PhD · Oncology, Genomics & Drug Development
AI-generated summary of the cited source, checked by automated accuracy review.
How we work
Share
Key Takeaway
Consider deep learning MRI models for TACE response in hepatocellular carcinoma, noting observational limitations.
This retrospective cohort study included 160 patients with Barcelona Clinic Liver Cancer (BCLC) stage A or B hepatocellular carcinoma greater than 3 cm. The population was divided into a training set of 112 patients, a test set of 48 patients, an independent cohort of 38 patients, and a follow-up cohort of 117 patients.
Researchers assessed AI models based on gadoxetic acid–enhanced MRI, including radiomics, deep learning, and clinical models, to predict treatment response and prognosis following transarterial chemoembolization. Independent clinical predictors identified included BCLC stage (P = 0.035), tumor number (P = 0.015), and tumor size (P = 0.013).
In the training set, area under the curve values were 0.79 for the clinical model, 0.84 for radiomics (XGBoost), and 0.96 for deep learning (DCNN). Test set results showed values of 0.70, 0.80, and 0.92 respectively. External validation sets yielded AUCs of 0.77, 0.80, and 0.86.
The deep learning model output served as an independent risk factor for overall survival with a hazard ratio of 15.9 (95% CI: 4.49-56.33; p < 0.001). Safety data regarding adverse events were not reported. Follow-up duration was not reported.
The study suggests these models may serve as a reliable tool for individualized treatment planning in precision oncology. However, the observational design limits causal inference regarding model efficacy in clinical practice.
Researchers analyzed data from patients with hepatocellular carcinoma to see if computer models could predict treatment outcomes. The study focused on people with Barcelona Clinic Liver Cancer stage A or B tumors larger than 3 centimeters. These patients were scheduled for transarterial chemoembolization, a common local therapy for liver cancer.
The team compared standard clinical information against AI models that used gadoxetic acid-enhanced MRI scans. The AI models applied radiomics and deep learning techniques to the images. They found that the deep learning model performed best, correctly predicting outcomes in 96% of the training group and 92% of the test group. Standard clinical scores were less accurate, reaching only 79% and 70% accuracy respectively.
The study also found that the AI model's output was linked to overall survival. Patients with a higher risk score from the model had a significantly lower chance of long-term survival. While the AI tools showed promise for planning individualized treatment, this research was based on a specific group of patients and the models were not yet proven for general use.
What this means for you:
AI models using MRI scans predicted liver cancer treatment response better than standard clinical scores in this study.
ObjectivesTo develop and validate radiomics and deep learning models based on the hepatobiliary phase (HBP) of gadoxetic acid-enhanced MRI (EOB-MRI) for the noninvasive prediction of treatment response and prognosis following transarterial chemoembolization (TACE) in hepatocellular carcinoma (HCC).Materials and methodsFrom April 2018 to September 2024, 160 patients with Barcelona Clinic Liver Cancer (BCLC) stage A or B HCC (>3 cm) were retrospectively enrolled and randomly divided into training (n = 112) and test (n = 48) sets. An independent cohort of 38 HCC patients was used for external validation. Twenty-six radiomic features were extracted using LASSO to construct a machine learning model using eXtreme gradient boosting (XGBoost). A deep convolutional neural network (DCNN) based on the ResNet50 architecture was used to develop a deep learning model. A clinical model was built via multivariate logistic regression. Model performance was evaluated by the area under the curve (AUC), calibration curves, and decision curve analysis (DCA). Kaplan–Meier analysis of combined deep learning and radiomics(DLR) scores was used to estimate overall survival in the follow-up cohort (n = 117).ResultsBCLC stage (P = 0.035), tumor number (P = 0.015), and tumor size (P = 0.013) were independent clinical predictors. In the training set, the AUCs (95% CI) for the clinical, radiomics (XGBoost), and deep learning (DCNN) models were 0.79, 0.84, and 0.96, respectively. In the test set, the AUCs were 0.70, 0.80, and 0.92, respectively. In the external validation set, the AUCs were 0.77, 0.80, and 0.86, respectively. The DCNN model showed superior calibration and the highest net clinical benefit in DCA. In addition, multivariable Cox regression revealed that DLR model output was an independent risk factor for the overall survival (hazard ratio: 15.9, 95% CI: 4.49-56.33; p < 0.001).ConclusionHBP-based AI models effectively predicted TACE response and prognosis in HCC patients, with the DCNN model showing the best performance. The integrated DLR model demonstrated high predictive accuracy and may serve as a reliable tool for individualized treatment planning in precision oncology.
