Predicting response to hypomethylating agents in myeloid neoplasms: a robust model

Area of Science:

• Hematology
• Molecular Biology
• Genomics

Background:

• Current biomarkers lack predictive power for hypomethylating agent (HMA) efficacy in myeloid neoplasms.
• There is a need for improved prognostic systems to guide HMA treatment decisions.

Purpose of the Study:

• To develop a novel predictive model for HMA response in myeloid neoplasms.
• To identify gene expression signatures associated with HMA treatment outcomes.

Main Methods:

• Weighted Gene Co-expression Network Analysis (WGCNA) to identify gene modules correlated with HMA response.
• Least Absolute Shrinkage and Selection Operator (LASSO) regression to build the HMA-29 predictive model using 29 genes.
• Validation across independent cohorts, disease subtypes, HMA drugs, and sampling conditions.

Main Results:

• The HMA-29 model stratified patients into responsive-like and resistant-like groups with high accuracy (AUC 0.9982).
• HMA-29 risk scores significantly predicted HMA response and patient survival (DFS and OS).
• Gene Set Enrichment Analysis (GSEA) revealed associations between HMA-29 risk and cell cycle pathways (G2M checkpoint, MYC, E2F targets).

Conclusions:

• The HMA-29 model demonstrates superior predictive and prognostic value for HMA response in myeloid neoplasms compared to existing systems.
• Findings suggest potential therapeutic strategies involving targeted therapies combined with HMA treatment.
• The model's stability across various conditions highlights its clinical utility.