Skip to content

Survival Analysis#

Survival analysis is a key statistical approach used to study time-to-event outcomes like time until disease progression, relapse, or death. Unlike traditional regression methods, survival analysis accounts for censoring (patients lost to follow-up or still alive at study end) and allows estimation of event probabilities over time.

Survival analysis is often used to:

  • Evaluate the impact of potential biomarkers on treatment response and patient outcomes.

  • Compare survival curves across patient subgroups (e.g., different mutation profiles, treatments, methods in multicentre studies, etc.).

  • Identify predictive markers of drug benefit or toxicity.

  • Build models that can use multiomic data to guide precision medicine.

Commonly used methods include:

  • Kaplan–Meier curves: non-parametric estimates of survival probabilities over time.

  • Cox proportional hazards models: regression framework to evaluate covariates (e.g., gene mutations, drug classes) affecting hazard rates.

  • Random survival forests / survival support vector machines / deep learning survival models: machine learning approaches that can capture complex interactions in high-dimensional omics datasets.

Recommended Learning Resources