Background and Aims: Metabolic dysfunction associated steatotic liver disease (MASLD) can eventually progress to hepatocellular carcinoma (HCC). Thirty to 40% of MASLD-HCC occur in the absence of fibrosis, posing a challenge for early detection. Previously, we reported 2 phenotypes of MASLD-HCC based on the severity of fibrosis (F0F1 vs. F3F4) (Buchard et al., 2021). The objective of this study is to explore lipid metabolism pathways through "omics" and identify biomarkers of MASLD-HCC based on the severity of fibrosis. Method: Our cohort included 56 pairs of tumor and non-tumor (NTT) human liver tissues (F0F1 = 28, F3F4 = 28), and 5 healthy liver tissues (HLT) used as control (CRB Foie). A non-targeted metabolomic analysis by liquid chromatography-mass spectrometry (LC-MS) was performed. In addition, a qRT-PCR analysis focused on a panel of 16 genes involved in the main lipid metabolism pathways was conducted (SPTLC1, SPTLC2, SGMS1, SGMS2, SMPD1, SMPD3, CHKA, PCYT1A, CHPT1, PEMT, MBOAT7, DGAT1, DGAT2, ACAT2, AADACL1 and ACOX1). For each group, TT was compared to NTT and to HLT. The study was approved by the Ethic Committee Sud-Est VI Clermont-Ferrand (Agreement number AU887, 04/03/2011). Results: First, using LC-MS, we identified 130 metabolites and 30 of them exhibited significant difference between TT and NTT. In the MASLD-HCC-F0F1, the levels of ceramide Cer(d18:1/22:0), sphingomyelin SM(41:1) and phosphatidylethanolamine PE(O-16:1_20:4) were decreased. While the levels of SM(35:1), SM(40:2), SM(42:3), phosphatidylcholines PC(16:0_14:0), PC(16:0/16:0), PC(18:0_18:1), PC(40:8) and PE(18:0_18:1) were increased. In contrast, most of metabolites content in MASLD-HCC-F3F4 were reduced compared to NTT, indicating an amplified anabolism. However, the levels of PC(18:0_18:1), PC(18:0_20:3) and phosphatidylinositol PI(16:0_18:1) were increased. Notably, PI(16:0_18:1) is considered as an onco-metabolite (Yang et al., 2018). Second, transcriptomic data revealed that in MASLD-HCC-F0F1, among the 16 genes studied, the expression of 6 genes coding for CHKA, DGAT1, MBOAT7, SPTLC2, SMPD1 and SGMS1 was up-regulated. However, MASLD-HCC-F3F4 shown no variation in the expression of these genes. Altogether, this data supports the existence of 2 MASLD-HCC phenotypes according to fibrosis severity. Furthermore, the increased expression of CHKA in MASLD-HCC-F0F1 is in line with our previous metabolomic data, reporting a significant accumulation of phosphocholine derivatives in MASLD-HCC-F0F1. This observation suggests that choline may serve as an imaging marker to detect HCC-F0F1. Conclusion: "Omics" analyses allow: 1) discrimination of MASLD-HCC based on fibrosis severity; 2) propose CHKA as a marker for MASLD-HCC-F0F1, which could lead to a clinical application: labeled choline may be favored as a more effective tracer for PET scanning in HCC patients with minimal fibrosis.