Statistical Validation of Surrogate Markers in Clinical Trials

Abstract

The increasing cost of drug development has raised the demand on the use of biomarkers as surrogate endpoints for the evaluation of new drugs in clinical trials. However, failed past attempts to use surrogate endpoints made it clear that, before deciding on the use of a candidate surrogate endpoint, it is of the utmost importance to investigate its validity. Such validation process has proven challenging for conceptual and practical reasons. In the present chapter, some of the statistical methods introduced for the evaluation of surrogate markers will be discussed. Emphasis will be made on the so-called meta-analytic approach and its information-theoretic version, where information from several units is combined to carry out the validation exercise. The methods will be illustrated using a case study in ophthalmology.