Interpretable multimodal deep learning for time-resolved survival prediction after hepatocellular carcinoma resection
10 affiliations
- 1Department of Radiology, The Third Hospital of Mianyang, Sichuan Mental Health Center, Mianyang, China
- 2Department of Radiology, The Third Affiliated Hospital of Shenzhen University (Luohu Hospital Group), Shenzhen, China
- 3Department of General Surgery, The First Affiliated Hospital of USTC, Division of Life Science and Medicine, University of Science and Technology of China, Hefei, China
- 4Department of Interventional Medicine Center, The Second People’s Hospital of YiBin, Yibin, China
- 5Zhejiang Key Laboratory of Multi-omics Precision Diagnosis and Treatment of Liver Diseases, Department of General Surgery, Sir Run-Run Shaw Hospital, Zhejiang University School of Medicine, Hangzhou, China
- 6Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
- 7Department of Radiology, Huzhou Central Hospital, Affiliated to Huzhou University, Huzhou, China
- 8Department of Radiology, West China Second Hospital, Sichuan University, Chengdu, China
- 9Department of Radiology, Luzhou Hospital of Traditional Chinese Medicine, Luzhou, China
- 10Department of Radiology, General Hospital of Ningxia Medical University, Yinchuan, China
Abstract
Hepatocellular carcinoma (HCC) exhibits substantial interpatient heterogeneity, leading to markedly variable outcomes and survival even among patients with similar stages and imaging phenotypes. Mainstream staging systems remain suboptimal, whereas pathology-dependent factors and high-cost genomic assays are neither scalable nor timely for clinical decision-making. Existing algorithms provide coarse risk stratification or static binary predictions, failing to capture the time-varying risk of death. We developed and externally validated TEMPO-HCC, a multimodal deep survival model with hierarchical interpretability, to estimate individualized overall survival risk trajectories after curative-intent resection. We curated a six-center cohort of 1475 patients and integrated multiphasic MRI, postoperative H&E whole-slide images, and perioperative predictors. A discrete-time survival head generated probabilities at 1, 2, 3, and 5 years after surgery. TEMPO-HCC outperformed unimodal models and guideline-based staging systems, achieving C-index of 0.751 and time-dependent AUCs of 0.836, 0.781, 0.812, and 0.680 at 12, 24, 36, and 60 months, respectively, in the external validation cohort. TEMPO-HCC represents a paradigm shift from static binary classification toward clinically actionable temporal risk prediction. Its hierarchical interpretability provides an auditable evidence chain linking macro-scale radiologic phenotypes to micro-scale histopathologic patterns. Importantly, augmenting guideline staging with TEMPO-HCC improved discrimination, enabling personalized postoperative surveillance and risk-adapted clinical management.
Related
- Interpretable CT-Based Radiomics for Prediction of High-Affinity Neoantigens Associated Immunogenic States in Advanced Hepatocellular Carcinoma Treated with TACE Plus Targeted Therapy and Immunotherapy.
- MIFF-CLIP-HCC: Contrastive Language-Image Pretraining-enhanced multimodal information fusion framework for early recurrence prediction in postoperative hepatocellular carcinoma.
- Preoperative prediction of microvascular invasion in hepatocellular carcinoma using routine clinical variables: development and temporal validation of an interpretable random forest model in a single-center retrospective cohort study.
- Peri-interventional neutrophil extracellular traps and immune surveillance in hepatocellular carcinoma: mechanisms, biomarkers, and recurrence after liver resection or thermal ablation
- Association of steatotic liver disease with long-term oncological outcomes after hepatic resection for hepatocellular carcinoma
BibTeX
@article{li2026interpretable,
title = {{Interpretable multimodal deep learning for time-resolved survival prediction after hepatocellular carcinoma resection}},
author = {Fan Li and Huancheng Yang and Ruishan Liu and Rundong Wang and Xu Feng and Yangyang Xie and Lian Yang and Wei Zhou and Xia Li and Xiang Wei and Dingbin Yuan and Die Hu and Hao Zhang and Jiahui Zhang and Yuxia Nie and Haibo Qu and Fei Wang and Jing Jia and Gang Ning},
journal = {NPJ Digital Medicine},
year = {2026},
month = jul,
doi = {10.1038/s41746-026-03027-0},
url = {https://doi.org/10.1038/s41746-026-03027-0},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Interpretable multimodal deep learning for time-resolved survival prediction after hepatocellular carcinoma resection
AU - Fan Li
AU - Huancheng Yang
AU - Ruishan Liu
AU - Rundong Wang
AU - Xu Feng
AU - Yangyang Xie
AU - Lian Yang
AU - Wei Zhou
AU - Xia Li
AU - Xiang Wei
AU - Dingbin Yuan
AU - Die Hu
AU - Hao Zhang
AU - Jiahui Zhang
AU - Yuxia Nie
AU - Haibo Qu
AU - Fei Wang
AU - Jing Jia
AU - Gang Ning
PY - 2026
DA - 2026/07/20
JO - NPJ Digital Medicine
DO - 10.1038/s41746-026-03027-0
UR - https://doi.org/10.1038/s41746-026-03027-0
AB - Hepatocellular carcinoma (HCC) exhibits substantial interpatient heterogeneity, leading to markedly variable outcomes and survival even among patients with similar stages and imaging phenotypes. Mainstream staging systems remain suboptimal, whereas pathology-dependent factors and high-cost genomic assays are neither scalable nor timely for clinical decision-making. Existing algorithms provide coarse risk stratification or static binary predictions, failing to capture the time-varying risk of death. We developed and externally validated TEMPO-HCC, a multimodal deep survival model with hierarchical interpretability, to estimate individualized overall survival risk trajectories after curative-intent resection. We curated a six-center cohort of 1475 patients and integrated multiphasic MRI, postoperative H&E whole-slide images, and perioperative predictors. A discrete-time survival head generated probabilities at 1, 2, 3, and 5 years after surgery. TEMPO-HCC outperformed unimodal models and guideline-based staging systems, achieving C-index of 0.751 and time-dependent AUCs of 0.836, 0.781, 0.812, and 0.680 at 12, 24, 36, and 60 months, respectively, in the external validation cohort. TEMPO-HCC represents a paradigm shift from static binary classification toward clinically actionable temporal risk prediction. Its hierarchical interpretability provides an auditable evidence chain linking macro-scale radiologic phenotypes to micro-scale histopathologic patterns. Importantly, augmenting guideline staging with TEMPO-HCC improved discrimination, enabling personalized postoperative surveillance and risk-adapted clinical management.
SN - 2398-6352
ER -