Preoperative MRI Intratumoral Heterogeneity Radiomics for Microvascular Invasion and Recurrence-Free Survival in HCC
Ruishan Liu1, Yangyang Xie2, Lv Yue1
1Department of Radiology, The Third Hospital of Mianyang, Sichuan Mental Health Center, Mianyang, China.
Summary
A new radiomics biomarker accurately predicts microvascular invasion (MVI) in hepatocellular carcinoma (HCC) before surgery. This noninvasive tool aids in stratifying recurrence risk, improving patient outcomes.
Area of Science:
- Hepatocellular Carcinoma Research
- Medical Imaging and Radiomics
- Oncology Biomarkers
Background:
- Microvascular invasion (MVI) is critical for hepatocellular carcinoma (HCC) recurrence and prognosis.
- Preoperative identification of MVI is challenging due to invasive pathology and sampling bias.
- Accurate MVI assessment is crucial for surgical planning and risk stratification.
Purpose of the Study:
- To develop and validate a noninvasive radiomics biomarker for preoperative MVI prediction in HCC.
- To assess the biomarker's ability to stratify postoperative recurrence-free survival risk.
- To compare the radiomics biomarker's performance against established staging systems.
Main Methods:
- Retrospective analysis of 567 HCC patients across four centers.
- Development of radiomics models using pretreatment MRI (global tumor region and intratumoral heterogeneity) and clinical data.
- Integration into a Fusion model for MVI prediction and prognostic stratification.
- Validation using AUC, calibration curves, decision-curve analysis, C-index, and time-dependent ROC analysis.
Main Results:
- The Fusion radiomics model demonstrated high discrimination for MVI prediction (AUCs 0.924 and 0.895 in validation and test sets).
- Ablation analyses confirmed incremental value from global tumor region, intratumoral heterogeneity, and clinical features.
- The Fusion score effectively stratified patients into high- and low-risk groups for recurrence-free survival.
- The prognostic model outperformed BCLC, CNLC, and AJCC staging systems in discrimination.
Conclusions:
- A preoperative radiomics biomarker integrating tumor phenotype and heterogeneity offers noninvasive MVI assessment in HCC.
- This biomarker can complement morphology-based staging for improved recurrence risk stratification.
- Radiomics provides a promising noninvasive approach for HCC management and personalized treatment strategies.
