Parametric survival analysis using R: Illustration with lung cancer data

Area of Science:

• Oncology
• Biostatistics
• Survival Analysis

Background:

• Cox regression is a prevalent survival model in oncology.
• Parametric survival models offer an alternative to Cox regression.
• This study applies semiparametric and various parametric models to lung cancer data.

Purpose of the Study:

• To illustrate the application of parametric survival models in lung cancer research.
• To compare the performance of different parametric models against Cox regression.
• To identify predictors of overall survival in African American lung cancer patients.

Main Methods:

• Utilized lung cancer data from 66 African American patients.
• Included patient stage, sex, age, smoking history, and tumor grade as predictors.
• Employed R software with the "Survival" package for analysis, comparing models using Akaike Information Criterion (AIC).

Main Results:

• Parametric models were fitted, controlling for age and stage.
• The log-logistic model exhibited the minimum AIC value (462.4087).
• The log-logistic model was determined to be the best fit for the studied African American lung cancer data.

Conclusions:

• Parametric survival models can be challenging to implement in routine cancer research.
• This paper provides a practical illustration of applying parametric survival models using R software.
• The study aims to facilitate the adoption of parametric survival models in clinical research.