Abstract
Aim: Acute myeloid leukemia (AML) is an aggressive hematological malignancy. Conventional risk stratification in AML fails to predict patient responses to targeted therapies such as Venetoclax, hindering precision medicine and the development of strategies to overcome drug resistance. Methods: We established an integrated multi-omics framework incorporating messenger RNA (mRNA)/long non-coding RNA (lncRNA) expression, DNA methylation, copy number alterations, and somatic mutation data. Using nine complementary clustering algorithms, we identified molecular subtypes in a discovery cohort and validated them in independent external cohorts. The multi-omics classification subsequently guided the screening of high-risk subtypes specific sensitizers to Venetoclax, with candidate efficacy validated through in vitro and in vivo experiments. Results: We identified three molecularly distinct AML subtypes with unique clinical features, a classification that was subsequently validated in independent external cohorts. Cluster 2 demonstrated the most favorable prognosis, while Cluster 3, characterized by high tumor protein 53 (TP53) mutation frequency and significant immune infiltration, exhibited the poorest outcomes and pronounced resistance to Venetoclax. Multi-omics-guided drug screening revealed that Cluster 3 displays particular sensitivity to both Elesclomol and the clinically available proteasome inhibitor Bortezomib. Through comprehensive in vitro and in vivo validation, we demonstrated that both agents significantly enhance the therapeutic efficacy of Venetoclax. Of particular translational relevance, the Venetoclax-Bortezomib combination leverages an agent with an established safety profile, offering a readily implementable strategy for near-future clinical translation. Conclusion: Our study establishes a novel multi-omics classification system that provides a robust foundation for investigating biological heterogeneity, elucidating resistance mechanisms, and developing effective combination strategies to achieve personalized therapy in AML.