Multi-omics cohort-based prediction model for early relapse of hepatocellular carcinoma post-surgery.

Abstract

3137 Background: Postoperative early relapse (PER) of hepatocellular carcinoma (HCC) presents significant challenges in clinical management. Identifying reliable predictive markers and therapeutic strategies for PER is crucial for improving patient outcomes. Methods: We constructed a predictive model for PER using multi-omics data from 177 HCC patients with follow-up information. Transcriptomic and proteomic profiling of HCC tissues was performed, followed by differential expression analysis and Weighted Gene Co-expression Network Analysis (WGCNA) to identify molecular markers associated with PER. Univariate, LASSO, and multivariate Cox regression analyses were employed to refine the marker set, resulting in a three-gene signature. The model's accuracy was validated using a proteomic cohort and The Cancer Genome Atlas (TCGA) database. Functional enrichment, drug sensitivity, and immune infiltration analyses were conducted to explore the biological characteristics and therapeutic implications of high-PER risk patients. Patient-derived organoid (PDO) models were used for further validation. Results: We identified 31 molecular markers associated with PER, which were narrowed down to a robust three-gene signature (MIK67, GPD1, and MBL2) with an area under the curve (AUC) of 0.868 for predicting early relapse. Functional enrichment analysis revealed that high-PER risk patients exhibited enhanced DNA damage repair and cell cycle pathways. Drug sensitivity analysis suggested potential benefits from gemcitabine and paclitaxel, which were validated using PDO models. Immune infiltration analysis showed reduced NK cell and M2 macrophage infiltration in high-PER risk patients, confirmed by single-cell sequencing and immunohistochemical validation. Conclusions: This study provides a novel multi-omics-based predictive model for early recurrence in HCC, highlighting potential therapeutic options for high-risk patients. The findings underscore the importance of DNA damage repair and cell cycle pathways in PER and suggest targeted therapies that could improve clinical outcomes for high-PER risk patients.