Integrating Magnetic Resonance Imaging-Based Deep Learning and Radiomics to Predict Treatment Response in Hepatocellular Carcinoma Receiving Combined Systemic and Locoregional Therapies: A Multicenter Study.
12 affiliations
- 1Department of Hepatobiliary Surgery and Liver Transplantation, Liver Cancer Institute and Zhongshan Hospital, Fudan University, Shanghai, China.
- 2Department of Research and Development, Shanghai United Imaging Intelligence Co., Ltd., Shanghai, China.
- 3Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai Institute of Medical Imaging, Shanghai, China.
- 4Department of Liver Surgery, Sun Yat-sen University Cancer Center, Guangzhou, China.
- 5Department of Hepatic Surgery (II), Eastern Hepatobiliary Surgery Hospital, Naval Medical University (Second Military Medical University), Shanghai, PR China.
- 6Department of Hepatic Oncology, Liver Cancer Institute and Zhongshan Hospital, Fudan University, Shanghai, China.
- 7Department of Hepatobiliary Surgery, The First Affiliated Hospital of Guangxi Medical University, Nanning, China.
- 8Department of Hepatobiliary, National Clinical Research Center of Cancer, Oncology Key Laboratory of Cancer Prevention and Therapy, Tianjin Medical University Cancer Insitute and Hospital, Tianjin, China.
- 9Department of Hepatobiliary Pancreatic Surgery, Fujian Provincial Hospital, Fuzhou, China.
- 10Department of General Surgery, Xiangya Hospital, Central South University, Changsha, China.
- 11Department of Hepatobiliary Surgery, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China.
- 12Department of Hepatobiliary Surgery, Affiliated Nantong Hospital 3 of Nantong University, Nantong, China.
Abstract
Background
Although combining systemic and locoregional therapies has improved outcomes for hepatocellular carcinoma (HCC), identifying appropriate candidates remains challenging because no reliable tool predicts response.Methods
This retrospective multicenter study enrolled 334 HCC patients who received first-line locoregional-systemic therapy from nine centers in China. Following automated tumor segmentation using nnU-Net on pretreatment contrast-enhanced magnetic resonance imaging, predefined radiomics features were computationally extracted, and deep learning features were obtained via a ResNet convolutional neural network. Three types of predictive models for tumor response (radiomics, deep learning, and combined) were developed and validated on the training cohort (n = 213), internal validation cohort (IVC, n = 53), and external validation cohort (EVC, n = 68). Performance was assessed using area under the receiver operating characteristic curve (AUC) and net reclassification improvement (NRI). Associations of model with progression-free survival (PFS) and overall survival (OS) were evaluated.Results
Among the three models (multi-sequence radiomics, multi-sequence deep learning, and combined), the combined model showed the best predictive performance in both the training (AUC: 0.997, 95% CI: 0.994-1.000) and validation cohorts (AUC of IVC 0.900, 95% CI: 0.806-0.994; AUC of EVC: 0.871, 95% CI: 0.790-0.951). The combined model demonstrated significantly improved predictive performance compared with the multi-sequence radiomic model (training NRI = 43.9%, p < 0.001; IVC NRI = 38.7%, p = 0.038; EVC NRI = 3.2%, p = 0.806) and multi-sequence deep learning model (training NRI = 80.8%, IVC NRI = 72.7%, EVC NRI = 50.4%; all p < 0.05). Specifically, the combined model was significantly associated with both PFS and OS.Conclusions
The combined radiomics-deep learning model outperformed both single-modality models in predicting objective response and stratifying survival outcomes in unresectable HCC patients receiving combined systemic and locoregional therapy.
Related
- Magnetic resonance imaging-based radiomics for predicting P53-mutated hepatocellular carcinoma: intratumoral and peritumoral analysis
- The Efficacy and Safety of Combined Locoregional and Systemic Therapy versus Systemic Therapy Alone in Patients with Hepatocellular Carcinoma and Portal Vein Tumor Thrombosis: A Multicenter Real-World Retrospective Cohort Study.
- CT habitat and peritumoral radiomics for predicting progression after repeat TACE in hepatocellular carcinoma: a multicenter study
- A multimodal deep learning radiomics nomogram integrating clinicopathological features for predicting pathological complete response after neoadjuvant therapy in triple-negative breast cancer: a multicenter retrospective study
- Post-transplant oncologic outcomes after locoregional downstaging before living donor liver transplantation for hepatocellular carcinoma beyond Milan criteria: a systematic review and meta-analysis of selected cohorts
BibTeX
@article{wang2026integrating,
title = {{Integrating Magnetic Resonance Imaging-Based Deep Learning and Radiomics to Predict Treatment Response in Hepatocellular Carcinoma Receiving Combined Systemic and Locoregional Therapies: A Multicenter Study.}},
author = {Lu-Na Wang and Shi-Qi Zhou and Fei Li and Li-Fang Wu and Li Xu and Kui Wang and Ning-Ling Ge and Yi Chen and Tao Peng and Tian-Qiang Song and Mao-Lin Yan and Lan Zhang and Le-Du Zhou and Yong-Jun Chen and Guo-Dong Zhu and Yang-Xun Pan and Hui Xue and Guang-Zhi Zhu and Dong-Ming Liu and Zhen-Xin Zeng and Kuan Hu and Yu-Chen Yang and Zi-Yue Yang and Tian He and Zi-Yi Wang and Cheng Huang and Ying-Hao Shen and Xiao-Dong Zhu and Hui Li and Qiang Gao and Jian Zhou and Jia Fan and Bin Xu and Hui-Chuan Sun and China Liver Cancer Study Group Young Investigators (CLEAP)},
journal = {Liver cancer},
year = {2026},
month = sep,
doi = {10.1159/lic/admag009},
url = {https://europepmc.org/article/MED/42924067},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - Integrating Magnetic Resonance Imaging-Based Deep Learning and Radiomics to Predict Treatment Response in Hepatocellular Carcinoma Receiving Combined Systemic and Locoregional Therapies: A Multicenter Study.
AU - Lu-Na Wang
AU - Shi-Qi Zhou
AU - Fei Li
AU - Li-Fang Wu
AU - Li Xu
AU - Kui Wang
AU - Ning-Ling Ge
AU - Yi Chen
AU - Tao Peng
AU - Tian-Qiang Song
AU - Mao-Lin Yan
AU - Lan Zhang
AU - Le-Du Zhou
AU - Yong-Jun Chen
AU - Guo-Dong Zhu
AU - Yang-Xun Pan
AU - Hui Xue
AU - Guang-Zhi Zhu
AU - Dong-Ming Liu
AU - Zhen-Xin Zeng
AU - Kuan Hu
AU - Yu-Chen Yang
AU - Zi-Yue Yang
AU - Tian He
AU - Zi-Yi Wang
AU - Cheng Huang
AU - Ying-Hao Shen
AU - Xiao-Dong Zhu
AU - Hui Li
AU - Qiang Gao
AU - Jian Zhou
AU - Jia Fan
AU - Bin Xu
AU - Hui-Chuan Sun
AU - China Liver Cancer Study Group Young Investigators (CLEAP)
PY - 2026
DA - 2026/09/07
JO - Liver cancer
DO - 10.1159/lic/admag009
UR - https://europepmc.org/article/MED/42924067
AB - Background Although combining systemic and locoregional therapies has improved outcomes for hepatocellular carcinoma (HCC), identifying appropriate candidates remains challenging because no reliable tool predicts response.Methods This retrospective multicenter study enrolled 334 HCC patients who received first-line locoregional-systemic therapy from nine centers in China. Following automated tumor segmentation using nnU-Net on pretreatment contrast-enhanced magnetic resonance imaging, predefined radiomics features were computationally extracted, and deep learning features were obtained via a ResNet convolutional neural network. Three types of predictive models for tumor response (radiomics, deep learning, and combined) were developed and validated on the training cohort (n = 213), internal validation cohort (IVC, n = 53), and external validation cohort (EVC, n = 68). Performance was assessed using area under the receiver operating characteristic curve (AUC) and net reclassification improvement (NRI). Associations of model with progression-free survival (PFS) and overall survival (OS) were evaluated.Results Among the three models (multi-sequence radiomics, multi-sequence deep learning, and combined), the combined model showed the best predictive performance in both the training (AUC: 0.997, 95% CI: 0.994-1.000) and validation cohorts (AUC of IVC 0.900, 95% CI: 0.806-0.994; AUC of EVC: 0.871, 95% CI: 0.790-0.951). The combined model demonstrated significantly improved predictive performance compared with the multi-sequence radiomic model (training NRI = 43.9%, p < 0.001; IVC NRI = 38.7%, p = 0.038; EVC NRI = 3.2%, p = 0.806) and multi-sequence deep learning model (training NRI = 80.8%, IVC NRI = 72.7%, EVC NRI = 50.4%; all p < 0.05). Specifically, the combined model was significantly associated with both PFS and OS.Conclusions The combined radiomics-deep learning model outperformed both single-modality models in predicting objective response and stratifying survival outcomes in unresectable HCC patients receiving combined systemic and locoregional therapy.
SN - 2235-1795
SN - 1664-5553
ER -