Multi-omics integration defines an ECM-associated intratumoral heterogeneity signature enabling prognosis assessment and therapeutic stratification in hepatocellular carcinoma.

Intratumoral heterogeneity (ITH) is a fundamental driver of hepatocellular carcinoma (HCC) progression and therapeutic resistance. Here, we developed a multi-omics-derived ITH scoring system to improve patient stratification and guide precision therapy. Using somatic mutation data from 361 TCGA HCC patients, we quantified genomic heterogeneity via the Mutant-Allele Tumor Heterogeneity (MATH) score, while single-cell RNA sequencing (scRNA-seq) of 20 HCC samples (our cohort and GSE149614) enabled assessment of cellular diversity. Through robust rank aggregation (RRA) of differentially expressed genes (DEGs) from both bulk RNA-seq (MATH-stratified) and scRNA-seq (malignant cell heterogeneity-stratified) analyses, we established a transcriptome-based ITH signature. High ITH HCC patients demonstrated aggressive clinicopathological features, including elevated serum AFP levels and microvascular invasion, and exhibited significantly worse clinical outcomes. Mechanistic investigations revealed that high ITH tumors were associated with poor differentiation and extracellular matrix (ECM) remodeling, characterized by increased collagen deposition, matrix stiffness, and activation of mechanotransduction pathways. Notably, we identified FK506 binding protein 10 (FKBP10) as a key collagen-related biomarker linked to high ITH and poor prognosis. Moreover, it is potentially associated with tumor ECM remodeling. Clinically, the ITH signature was found to be correlated with diminished therapeutic efficacy of both conventional transarterial chemoembolization (TACE) and immune checkpoint blockade (ICB). Our study provides a clinically actionable framework for quantifying HCC heterogeneity and reveals ECM remodeling as a potential therapeutic target in high ITH tumors, offering new avenues for personalized treatment strategies.