Interpretable fusion of multiphase gadoxetic acid-enhanced MRI habitat radiomics for predicting extrahepatic metastasis after hepatectomy in hepatocellular carcinoma
Frontiers in Oncology · 20.11.2026 · 16
DOI: 10.3389/fonc.2026.1965659 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Oncology, online erschienen am 20.11.2026. Verfasst von Likai Wang und 5 weiteren Autor:innen. Die Arbeit nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Preoperative risk stratification for extrahepatic metastasis after hepatectomy remains challenging in patients with hepatocellular carcinoma. This study aimed to develop an interpretable fusion model integrating clinical variables, whole-tumor radiomics, and habitat radiomics derived from multiphase gadoxetic acid-enhanced MRI.
Methods
This retrospective study included 280 patients with hepatocellular carcinoma who underwent hepatectomy and were randomly divided into a training cohort (n = 196) and an internal test cohort (n = 84). Radiomics and habitat radiomics features were extracted from arterial, portal venous, and hepatobiliary phase images. Clinical, whole-tumor radiomics, habitat radiomics, and fusion models were developed in the training cohort, and model performance was evaluated in the test cohort. SHapley Additive exPlanations were used to interpret the fusion model.
Results
Ninety-six of 280 patients (34.3%) developed extrahepatic metastasis within 24 months after surgery. In the test cohort, the areas under the receiver operating characteristic curve were 0.814 for the clinical model, 0.742 for the whole-tumor radiomics model, 0.812 for the habitat radiomics model, and 0.878 for the fusion model. The fusion model’s test-cohort accuracy, sensitivity, and specificity were 78.6%, 75.9%, and 80.0%, respectively, with a Brier score of 0.1435.
Conclusion
The multiphase MRI-based fusion radiomics model demonstrated promising discrimination for EHM after hepatectomy in patients with HCC, suggesting that combining clinical information with different radiomics features may offer potential predictive value and provide supplementary information for preoperative EHM risk assessment.
Zitieren
APA
Wang, L., Zhao, K., Sun, W., Dong, Q., Han, F., & Li, T. (2026). Interpretable fusion of multiphase gadoxetic acid-enhanced MRI habitat radiomics for predicting extrahepatic metastasis after hepatectomy in hepatocellular carcinoma. Frontiers in Oncology, 16. https://doi.org/10.3389/fonc.2026.1965659
Vancouver
Wang L, Zhao K, Sun W, Dong Q, Han F, Li T. Interpretable fusion of multiphase gadoxetic acid-enhanced MRI habitat radiomics for predicting extrahepatic metastasis after hepatectomy in hepatocellular carcinoma. Frontiers in Oncology. 2026;16. doi:10.3389/fonc.2026.1965659
BibTeX
@article{Wang2026,
author = {Wang, Likai and Zhao, Kexin and Sun, Wenbo and Dong, Qinghua and Han, Fang and Li, Tianran},
title = {Interpretable fusion of multiphase gadoxetic acid-enhanced MRI habitat radiomics for predicting extrahepatic metastasis after hepatectomy in hepatocellular carcinoma},
journal = {Frontiers in Oncology},
year = {2026},
volume = {16},
doi = {10.3389/fonc.2026.1965659}
}
EUVIMED