Discovery of miRNA–RNA Biomarkers for Risk Stratification in Acute Myeloid Leukemia with Multi-Cohort Validation
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Acute myeloid leukemia (AML) is a highly heterogeneous malignancy, and current prognostic standards such as the European LeukemiaNet (ELN) do not fully capture its underlying regulatory complexity. We developed a two-step, PCA-based survival workflow that independently and jointly analyzes gene and miRNA expression to identify biomarkers for patient risk stratification, followed by support vector machine (SVM) validation across multiple AML cohorts. This approach yielded a 19-gene panel—including known oncogenes (HMGA2, TAL1) and novel candidates (MLEC, APOE)—that achieved high validation accuracy relative to ELN (AUCs > 0.879). Parallel analyses identified a 16-miRNA panel enriched for tumor suppressors (miR-7b-3p, miR-26a-5p) and novel markers (miR-3613-5p, miR-942-5p). Integrating experimentally supported miRNA:gene interactions revealed 10 coherent regulatory pairs, most showing inverse correlations, improving performance over single-omic models. Finally, a Cox regression–based risk score accurately stratified patients and outperformed ELN. Overall, this framework provides biologically grounded biomarkers with strong prognostic power.
Statement of significance
This work introduces a multi-omic miRNA:gene framework that models regulatory programs at genome-wide scale and links miRNAs to their functional targets to generate mechanistically interpretable biomarkers. The resulting biomarkers provide a risk score that improves upon ELN-based stratification, offering a more precise and biologically informed foundation for clinical risk prediction and patient classification.
