Hepatic carcinoma causes death mainly by disseminated metastasis progression from the organ being confined.¹ Prognostic diagnosis of different metastasis stages which are closely related to cellular metabolism is a major challenge.^2,3 Metabolomics analysis based on ¹H NMR spectroscopy could provide ability to quantify specific alterations related to metastasis potentials of different cell lines.^4-6We explored metabolic characteristics of normal hepatocyte and hepatocarcinoma cell lines related to different metastasis potentials with ultimate goal of determining the molecular mechanisms of metastasis.We have performed NMR-based metabolic analysis of normal hepatocyte LO2 and cell lines from hepatocarcinoma including lowly metastatic HepG2 and highly metastatic MHCC97L and MHCC97H. Combining with methods of principal component analysis and orthogonal projection to latent structure with discriminant analysis, cells with different metastasis potentials can be separated according to their metabolic profiles. We determined the characteristic metabolites with statistically significance level of P<0.05 or P<0.01 between normal and cancer cell lines (Fig. 1). There were several common characteristic metabolites between normal and lowly and between normal and highly metastatic cell lines, including valine, lactate, acetate, proline, phosphocholine, and glucose, but no characteristic metabolite was identified for highly metastatic cell lines (Fig. 2&3). The results showed that there was a significant difference in the glucose consumption of different metastatic cells. Higher metastatic cells tended to cause higher level of glycolysis as well as relative products of intermediates. Meanwhile, the cultured medium was also analyzed and the characteristic metabolites obtained from cultured medium were further confirmed by excluding the discriminary metabolites associated with pure medium. Four classifiers based on the medium-derived characteristic metabolites were established by support vector machines to identify normal and cancer cell lines and achieved great diagnostic sensitivities and specificities of >93% (Fig. 4). Specifically, to identify normal and highly metastatic cell lines, both of the obtained sensitivities and specificities reached 100%. Such results provided evidence that metabolic analysis of cultured medium could be a valid method to understand metabolic alterations associated with different metastatic cells.Our results demonstrate that ¹H NMR spectroscopy has potential use in metastasis progression of hepatic carcinoma and will be helpful for the determination of metabolic markers for hepatic carcinoma identification. We also notice the extension scope of the proposed characteristic metabolites between different tumors. Future work should be placed on this subject including the spread of proposed characteristic metabolites between different tumors in the potential clinical study.