What question did this study set out to answer?

The aim is to create a radiomics model using DCE-MRI to predict vascular patterns in hepatocellular carcinoma and assess their prognostic significance.

March 29, 2026Open Access

MRI Radiomics-Based Evaluation of Vessels Encapsulating Tumor Clusters and Microvascular Invasion in Hepatocellular Carcinoma

Key Points

The aim is to create a radiomics model using DCE-MRI to predict vascular patterns in hepatocellular carcinoma and assess their prognostic significance.
Retrospective analysis of 306 patients with hepatocellular carcinoma undergoing radical resection.
Patients were divided into training and validation cohorts across two medical centers.
Radiomics features were extracted from different phases of DCE-MRI images.
A combined clinical-radiomics model was developed to enhance prediction accuracy.
The combined model achieved AUC values of 0.857 in the training cohort, 0.761 in internal validation, and 0.723 in external validation.
Vascular pattern positivity was linked to early recurrence and shorter disease-free survival.
The model shows promising potential for noninvasive preoperative assessment.

Abstract

Purpose: This study aimed to develop a radiomics model based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) for the preoperative prediction of two distinct histopathological vascular patterns in hepatocellular carcinoma (HCC): vessels encapsulating tumor clusters (VETC) and microvascular invasion (MVI). In addition, the study evaluated the prognostic significance of these vascular patterns in predicting postoperative outcomes in patients with HCC. Patients and Methods: A total of 306 patients with HCC who underwent radical resection at two medical centers were retrospectively included. Patients from Center 1 were randomly assigned to a training cohort and an internal validation cohort at a ratio of 7:3, while patients from Center 2 comprised the external validation cohort. Radiomics features were extracted from the arterial phase (AP), portal venous phase (PP), and delayed phase (DP) DCE-MRI images, including intratumoral, peritumoral, and fused intra-peritumoral regions. Radiomics models were constructed based on these features, and the optimal model was subsequently integrated with clinical variables to establish a combined clinical–radiomics model. Results: For predicting VETC and/or MVI (defined as VM patterns), the combined model achieved area under the curve (AUC) values of 0.857 in the training cohort, 0.761 in the internal validation cohort, and 0.723 in the external validation cohort. Calibration and decision curve analysis (DCA) indicated acceptable calibration performance and potential clinical utility of the combined model. Both pathological VM positivity and model-predicted VM positivity were significantly associated with early recurrence (ER) and shorter disease-free survival (DFS). Conclusion: The clinical–radiomics combined model based on DCE-MRI holds potential value as a noninvasive preoperative approach for evaluating VM patterns and postoperative prognosis in patients with HCC. Keywords: hepatocellular carcinoma, radiomics, VETC, MVI, DCE-MRI, prognosis

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper