Deep Learning-Based Prediction of Transarterial Chemoembolization Target Lesion Response in Hepatocellular Carcinoma Post-TACE Using Pre- and Post-MRI Data

Tuesday, April 1, 2025

3:27 PM – 3:36 PM CT

Location: Music City Center, 209 ABC

Presenting Author(s)

Tej I. Mehta, MD (he/him/his)

Interventional Radiology Resident
Johns Hopkins Hospital, United States

Author/Co-author(s)

MA

Moustafa Abou Areda, MD

Radiology resident
Johns Hopkins University School of Medicine, United States
CH

Caleb Heiberger, MD

Radiology Resident
Johns Hopkins, United States
Tony Tae Kyung Kim, MD

Resident Physician
Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins Hospital, United States
RL

Robert P. Liddell, MD, FSIR

Division Chief, Associate Professor Radiology and Radiological Sciences
The Johns Hopkins Hospital, Baltimore, Maryland, United States, United States
HB

Harrison Bai, MD, MS

Director, Radiology AI Lab
Johns Hopkins University, United States

Purpose: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide. Transarterial chemoembolization (TACE) is a primary locoregional therapy for patients with intermediate-stage HCC. Early prediction of target lesion (TL) response to TACE, as defined by the modified Response Evaluation Criteria in Solid Tumors (mRECIST), is critical for optimizing patient management and improving outcomes. We sought to develop and validate a convolutional neural network (CNN) model utilizing pre- and post-TACE magnetic resonance imaging (MRI) data for early prediction of TL response by mRECIST criteria in patients with HCC.

Materials and Methods: 716 patients with biopsy-confirmed HCC who underwent TACE and had pre-TACE and 6 months post-TACE MRI scans. Segment Anything, a deep learning-based semi-automated segmentation algorithm, was used for semi-automated segmentation of target HCC lesions on arterial phase post-contrast T1, T2, and diffusion weighted sequences. The CNN model was pretrained on the Liver Tumor Segmentation (LiTS) Challenge dataset and fine-tuned using the segmented images from our cohort to characterize HCC lesions. The CNN architecture comprised multiple convolutional and pooling layers to extract hierarchical imaging features, followed by fully connected layers for classification into mRECIST categories: complete response, partial response, stable disease, or progressive disease. Transfer learning was employed by using pretrained weights from the LiTS dataset. The fine-tuning process involved training the CNN on our specific dataset to adapt the model to the nuances of our patient population and imaging protocols. The dataset was split into training (80%) and testing (20%) sets. Model performance was evaluated using accuracy (AC), sensitivity (SE), specificity (SP), and area under the receiver operating characteristic curve (AUC).

Results: The CNN model achieved an AUC of 0.93 for TL prediction by mRECIST criteria at 6 months post-TACE. It demonstrated an AC of 87.5%, SE of 89.5% and SP of 85.7%. The model correctly classified 126 of 143 cases in the test set.

Conclusion: The CNN-based model effectively predicts TL response by mRECIST criteria using pre- and post-TACE MRI data in patients with HCC. This approach could assist in prognostication of treatment outcomes post-TACE and tailor treatment strategies accordingly.