Background: Multiple myeloma (MM) is the second most common hematological malignancy, characterized by the abnormal accumulation of plasma cells in the bone marrow. Although the latest treatments, including proteasome inhibitors, immunomodulating agents, or immunotherapy, have greatly improved patient survival, a residual subset of cells remains resistant to therapies and usually causes relapses. Aims: The drug resistance could be explained, among others, by the interactions with the microenvironment, MM’s high molecular heterogeneity, or the appearance of adaptive survival mechanisms after treatment exposure. However, a better understanding of the mechanisms involved in drug resistance remains of significant interest. Among the factors influencing the resistance of cancer cells, the “metabolic plasticity” of the tumor and, therefore, its ability to adapt to stress conditions is a mechanism increasingly studied in recent years in cancer. Although measuring mitochondrial metabolism has been identified as a major factor influencing response to treatments in several cancers, few studies have been documented in MM. Methods: Here, we characterized the metabolic profiles of a panel of 20 human MM cell lines (HMCL) representative of the molecular heterogeneity found in MM patients. This panel included 10 commercial HMCL and 10 HMCL derived in our laboratory. First, oxygen consumption rates (OCR), extracellular acidification rate (ECAR), and spare respiratory capacity (SRC) of our HMCL panel were assessed in a Mito Stress Assay using a Seahorse XFe96 analyzer. Results: Interestingly, the metabolic activities were shown very heterogeneous in HMCLs with a part of cell lines more dependent of the glycolysis activity and another part more dependent of the mitochondrial respiration. By integrating the HMCL’s metabolic profiles with their respective transcriptomic data (RNAseq), we defined a metabolomic score to classify the HMCL into different groups and represent their glycolysis level or mitochondrial activity. The score was calculated from the expression of 112 genes involved in the electron transport chain (Oxphos) or in glycolysis. For a given gene, the expression values were first log2-transformed and then normalized by subtracting the average. The score is computed as the difference between the Oxphos and glycolytic genes’ average expression. Interestingly, high significant correlations between the HMCL’s functional metabolic profiles and their calculated metabolomic score were identified. Furthermore, we investigated the potential prognostic value of this gene-based metabolomic score in the MMRF CoMMpass cohort (newly diagnosed MM patients, n=674) using the maxstat algorithm, which segregated the cohort into two groups with a significantly different outcome. We identified that 32% of patients with a high gene-based metabolomic score, were associated with poor overall survival (P = 3.1x10-6). We then validated this prognostic value in a second cohort of MM’s patients: the cohort of 206 patients (ArrayExpress public database under accession number E-MTAB-362). Moreover, we performed correlation analyses between the metabolic profiles of those HMCL and their respective response to the conventional treatments (IC50). We found a significant correlation between a high mitochondrial ATP production and the resistance to proteasome inhibitor (P = 0.023, n= 12). Summary/Conclusion: Altogether, we demonstrated that metabolomic deregulation could participate in drug resistance in MM. Inhibitors targeting metabolic activities may be of therapeutic interest to overcome drug resistance in MM.