Building a Risk Scoring Model for ARDS in Lung Adenocarcinoma Patients Using Machine Learning Algorithms

Area of Science:

• Oncology
• Genomics
• Pulmonology

Background:

• Lung adenocarcinoma (LUAD) is a major non-small-cell lung cancer subtype.
• Acute respiratory distress syndrome (ARDS) complicates LUAD, increasing mortality.
• Understanding ARDS-related genes in LUAD is vital for improving patient outcomes.

Purpose of the Study:

• To identify ARDS-related genes influencing LUAD prognosis.
• To develop a predictive model for LUAD patient survival.
• To explore the functional and genetic characteristics of these genes in LUAD.

Main Methods:

• Utilized TCGA, GEO, and GTEx data for differential gene expression analysis.
• Employed univariate Cox regression, consensus clustering, and machine learning for model development.
• Conducted functional enrichment, immune infiltration, copy number variation, and mutational burden analyses, validated with single-cell data.

Main Results:

• Identified 276 ARDS-related genes significantly associated with LUAD prognosis.
• Developed a machine learning-based risk scoring model with accurate survival prediction capabilities.
• Found these genes are involved in cell cycle regulation and immune cell infiltration, with distinct genetic profiles.

Conclusions:

• ARDS-related genes hold significant prognostic value in LUAD.
• The developed risk model offers a tool for personalized therapy and prognosis in LUAD.
• Further research is warranted to validate findings and explore clinical applications of these biomarkers.