Surrogate Marker Evaluation: A Tutorial Using R

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Research

Background:

• Surrogate markers are widely used in clinical studies as alternatives to primary outcomes.
• Validation is crucial for surrogate markers to reliably assess treatment effectiveness.
• Existing statistical and clinical research has extensively explored surrogate marker evaluation over 35 years.

Purpose of the Study:

• To describe available statistical frameworks for evaluating surrogate markers.
• To focus on the practical implementation of the proportion of treatment effect explained (PTE) framework.
• To provide R code for implementing these procedures.

Main Methods:

• Review of statistical frameworks for surrogate marker evaluation.
• Focus on the PTE framework for both uncensored and censored outcomes.
• Inclusion of parametric and non-parametric estimation methods.
• Consideration of multiple surrogates, heterogeneity, and prediction perspectives.

Main Results:

• The tutorial details various statistical approaches for surrogate marker validation.
• Practical implementation guidance using R code is provided.
• Discussion includes advanced topics like the surrogate paradox and heterogeneity.

Conclusions:

• The PTE framework offers a valuable approach for surrogate marker evaluation.
• Further research is needed, particularly for using surrogate markers to test treatments in future studies.
• The study enriches the field with new insights and practical tools for surrogate marker analysis.