Background: In advanced hepatocellular carcinoma (HCC), combined treatment using anti-PD-L1 and anti-VEGF antibodies is a major therapeutic strategy in addition to using anti-PD-1 and anti-CTLA4 target agents. To identify clinically relevant molecular biomarkers for immunotherapeutic agents-based treatment in advanced HCC by comprehensive molecular profiling using genomics alterations and transcriptomic data. Methods: Multi-omics data from clinical tumor samples of patients treated with atezolizumab + bevacizumab (94 patients), nivolumab (40 patients), and ipilimumab + nivolumab (32 patients) were obtained from tumor biopsies or surgical resections in a single institution. The targeted panel sequencing was conducted via the Oncomine comprehensive assay (ThermoFisher, USA). QuantSeq mRNA sequencing (Lexogen, USA) was used to generate datasets for the genomic alterations and gene expression profiles. Results: A total of 166 target sequences for genomic alterations and 136 transcriptomic sequences were performed in patients with advanced HCC treated with immunotherapeutic agent-based treatment options. High tumor mutation burden was significantly associated with favorable therapeutic efficacy (disease control rate, P=0.036) and progression-free survival (P=0.037) in patients who had atezolizumab with bevacizumab treatment. The genomic alteration of CTNNB1 mutation correlated with the favorable therapeutic efficacy of nivolumab monotherapy (PFS, P=0.022), contrary to TP53 mutation (PFS, P=0.063) which demonstrated unfavorable outcomes. CD274(PD-L1) mRNA expression was significantly associated with a high drug response following nivolumab monotherapy (objective response rate, P=0.047) and oncologic outcomes (PFS, P=0.021). Interestingly, PLA2G4A, known as a downstream molecule of the VEGF signaling pathway, was highly correlated with unfavorable therapeutic efficacy of the combined atezolizumab with bevacizumab treatment (disease control rate, P=0.002) and related oncologic outcomes (PFS, P=0.003). We also identified the clinically relevant molecular subtypes from unbiased clustering using the ontologic pathway collection. We also used gene expression profiles to predict the therapeutic response and oncologic outcomes from immunotherapeutic agent-based treatment options. Conclusions: We identified clinically relevant molecular biomarkers from comprehensive clustering of a multi-omics dataset to predict the therapeutic efficacy and oncologic outcomes. Further prospective studies with clinical validation are warranted to confirm clinical implications.