Background: HCC is one of the leading causes of cancer deaths worldwide. Despite increasing systemic therapy options for patients with advanced HCC, there are currently no predictive biomarkers to guide therapy selection. Phosphoproteomics has been successfully used to identify accurate predictive biomarkers in other cancers from frozen clinical samples (Dokal et al., 2021 ASCO Annual Meeting). In this study, we test the feasibility of biomarker discovery using phospho- and whole-proteomics analyses on formalin fixed and paraffin embedded (FFPE) tru-cut liver biopsies from patients with HCC. Methods: FFPE tru-cut biopsies (n = 29) were obtained from separate cohorts of 15 HCC patients [tumor (T)] and 14 patients with chronic liver disease with varying levels of cirrhosis, fibrosis or steatosis but without cancer [non-tumor (NT)]. T samples were from advanced HCC patients with varied aetiology (chronic viral hepatitis, alcohol or NAFLD). Proteins were extracted from 10 x 10µm sections; crosslinks reversed before digestion to peptides; and multiple clean-up/enrichment steps before analysis. Peptides were quantified by mass spectrometry and performance assessed using multi-variate and enrichment analyses. Results: Overall, 4978 phosphopeptides and 3721 proteins were quantified across all samples. Multi-variate analyses based on the relative expression of phosphopeptides or proteins showed a clear separation between T and NT samples, as well as phenotype-distinct separation (i.e according to differentiation) within the T group. Subsequent testing also verified that the method could reproducibly measure these HCC-differentiating phosphorylation sites (and proteins). Proteins associated with HCC were regulated as expected in T vs NT, e.g. ASS1 (log2-fold change (FC) = -1.75, p < 0.001), BDH2 (log2-FC = -1.97, p < 0.001). Known drug targets were found to be significantly upregulated in poorly-differentiated HCC, e.g. PARP1 (log2-FC = 1.7, p < 0.001) and AKT1 (log2-FC = 1.5, p < 0.001), as well as previously described phosphorylation sites, e.g. p-CDK1 T14 (log2-FC = 5.95, p = 0.007) and p-MCM2 S139 (log2-FC = 2.98, p = 0.046). Kinase substrate enrichment analysis (Casado et al., Science Signaling 2013) showed increased activity for several kinases in T vs NT, including A-RAF (z-score = 3.2, p < 0.001), MAPK3 (z-score = 2.1, p = 0.022) and TNK2 (z-score = 2.1, p = 0.003). Conclusions: We successfully adapted (phospho-)proteomics to FFPE tru-cut biopsy specimens, enabling reproducible identification of novel and known (phospho-)proteins that distinguished differentiation status. A prospective study utilising this phosphoproteomic analysis pipeline is now underway, recruiting patients with advanced HCC receiving first line standard of care therapy with immune checkpoint inhibitor combinations or tyrosine kinase inhibitors to identify potential predictive biomarkers of response.