Molecular Characterization of the Highest Risk Adult Patients with Acute Myeloid Leukemia (AML) through Multi-omics Clustering
Abstract Background: Acute myeloid leukemia (AML) is a clinically heterogeneous group of diseases with poor outcomes that are partly due to its complex and poorly understood heterogeneity. Methods: Here, we use a multi-omics approach to identify a molecular subgroup with the worst response to chemotherapy, and to identify promising drug targets specifically for this AML subgroup. Results: Multi-omics clustering analysis using RNA and CNA expression data resulted in the three primary clusters of 166 AML adult TCGA cancer cases. One of these clusters, which we label as the high-risk molecular subgroup (HRMS), consisted of cases that responded very poorly to standard chemotherapy, with only about 10% survival to two years. The gene TP53 was mutated in most cases in the HRMS, but not in other cases. The top 6 genes over-expressed in the HRMS included E2F4, CD34, CD109, MN1, MMLT3, and CD200. Multi-omics pathway analysis using RNA and CNA expression data identified in the HRMS subgroup over-expressed pathways related to immune function, cell proliferation, and DNA damage. Conclusion: Some AML patients are not successfully treated with the current standard of care chemotherapy, and urgently need targeted therapeutics. Potential drug targets include over-expressed genes E2F4, and MN1, as well as mutations in TP53, and several molecular pathways.