Survival prediction in acute myeloid leukemia using gene expression profiling

Area of Science:

• Hematology
• Genomics
• Bioinformatics

Background:

• Acute myeloid leukemia (AML) is a complex blood cancer with poor survival rates.
• Genetic heterogeneity in AML poses challenges for accurate prognosis.
• Predictive models are needed to improve patient outcomes.

Purpose of the Study:

• To develop a machine learning model for precise prognosis prediction in AML.
• To identify key prognostic factors in AML patients.
• To stratify AML patients into distinct risk groups.

Main Methods:

• Prognosis-related genes were identified using Kaplan-Meier analysis.
• A random forest model was built using age, TP53 mutation, ELN classification, and gene expression.
• Gene set enrichment analysis and SNF-based clustering were performed.

Main Results:

• The random forest model achieved an AUC of 0.75.
• The risk score significantly correlated with shorter overall survival.
• Three distinct AML patient subgroups were identified, with one group showing poorer prognosis.

Conclusions:

• A random forest-based risk score provides an effective method for AML prognosis.
• This model can aid in clinical decision-making for AML patients.
• Further research can refine risk stratification and treatment strategies.