Acute myeloid leukemia risk stratification in younger and older patients through transcriptomic machine learning...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Acute Myeloid Leukemia (AML) is a complex disease with incidence and genetic alterations varying by age.
• Current AML classification systems do not fully integrate age as a defining factor.
• Understanding age-related molecular nuances is critical for personalized treatment.

Purpose of the Study:

• To investigate age-related transcriptomic differences in Acute Myeloid Leukemia (AML).
• To develop a machine learning model for accurate AML risk prediction incorporating age.
• To identify novel prognostic biomarkers and understand the biological context of age in AML.

Main Methods:

• Analysis of RNA-sequencing data from 404 de novo AML patients.
• Application of a k-mer-based machine learning approach for transcriptomic analysis.
• Identification of gene signatures and prognostic biomarkers associated with age and risk groups.

Main Results:

• A machine learning model achieved over 90% accuracy in AML risk prediction.
• Key gene signatures were identified, distinguishing favorable and adverse ELN2017 risk groups.
• Distinct tumor profiles and immune/stromal cell population differences were observed across age groups, especially in older patients.

Conclusions:

• Age is a significant factor influencing transcriptomic complexity and tumor profiles in AML.
• Age-related molecular features are important for refining AML risk stratification.
• These findings offer potential for developing novel, age-specific therapeutic targets in AML.