Review of statistical methods for survival analysis using genomic data

Area of Science:

• Biostatistics
• Genomics
• Machine Learning

Background:

• Survival analysis is crucial for modeling time-to-event data, often featuring censored observations.
• Traditional statistical methods are established but may struggle with high-dimensional 'omics' data.

Purpose of the Study:

• To review statistical methods for survival analysis, particularly for high-dimensional genomic data.
• To bridge the gap between traditional methods and modern machine learning approaches.

Main Methods:

• Review of traditional survival analysis techniques.
• Exploration of regularization methods with various penalty functions for genomic data.
• Description of machine learning adaptations for survival analysis.

Main Results:

• Traditional methods are effective for censored data but limited with high-dimensional 'omics' data.
• Regularization and machine learning offer advanced solutions for complex predictive survival modeling.
• These advanced methods improve prediction accuracy for individual survival probabilities.

Conclusions:

• A comprehensive understanding of survival analysis methods is vital for researchers.
• Integrating regularization and machine learning enhances the analysis of high-dimensional genomic survival data.
• This review aids clinicians and researchers in selecting appropriate statistical tools.