Identification of Microvascular Invasion-Related Biomarkers for Personalized Treatment of Hepatocellular Carcinoma.

Introduction: Hepatocellular Carcinoma (HCC) exhibits high recurrence rates, particularly when accompanied by Microvascular Invasion (MVI). We identified MVI-related biomarkers and established a prognostic model for personalized HCC treatment.

Methods: Data were downloaded from The Cancer Genome Atlas (TCGA) and HCCDB databases. Key radiomics features were identified using the support vector machine-recursive feature elimination (SVM-RFE) algorithm, and differential expression analysis was performed with DESeq2. This was followed by functional enrichment analysis using the clusterProfiler package. Through univariate and Lasso regression analyses, we constructed a robust RiskScore model to effectively stratify HCC patients into distinct risk groups based on the median RiskScore value. The model prediction performance was evaluated using ROC curves and Kaplan-Meier (KM) analysis. We used the CIBERSORT algorithm to characterize immune cell infiltration patterns and conducted GSEA to identify differentially activated pathways between the risk groups.

Results: Radiomic analysis revealed four significant features strongly associated with MVI, enabling the construction of a nomogram model with robust classification performance (AUC = 0.742). Subsequent analysis identified 241 overlapping MVI-related Differentially Expressed Genes (DEGs) enriched in critical tumor proliferation and invasion pathways. A 10-gene RiskScore model was developed, demonstrating excellent prognostic discrimination in training and validation cohorts. CIBERSORT analysis revealed significant correlations between specific immune cell infiltration and the 10 genes. GSEA analysis showed significant enrichment of cell cycle regulation pathways in the high-risk group, suggesting their important role in MVI.

Discussion: The RiskScore was established using MVI-related features for prognosis assessment in HCC.

Conclusion: Our findings provided novel biomarkers and a theoretical basis for the early diagnosis and personalized treatment of HCC.