Validation of surrogate endpoint biomarkers in prostate cancer chemoprevention trials

Phase 2 cancer chemoprevention trials are designed to provide estimates of the efficacy of an agent at a specified dose, and the expected size of the risk reduction that may be achieved in a subsequent phase 3 randomized trial. To allow these trials to be rapid and efficient, the study outcome is modulation of a surrogate endpoint biomarker (SEB), that is, a molecular event assumed to be on the causal pathway between the chemopreventive agent and the desired reduction in cancer incidence. However, SEBs commonly used in prostate cancer chemoprevention studies, such as prostate-specific antigen, high grade prostatic intraepithelial neoplasia, proliferation, and apoptosis, have not been validated by documenting that changes in the SEBs ultimately translate to decreased prostate cancer risk. Because of uncertainty in the pathway from SEBs to cancer, additional considerations are necessary to permit valid inferences from phase 2 trial data. This article considers the framework underlying validation and use of SEBs in specific chemoprevention models and methodologic issues in quantifying the effect of an agent. In particular, inferences depend on whether a single pathway involving the SEBs is assumed to mediate the effect of the agent on cancer incidence. If there are competing pathways of equal or greater importance than the one involving the candidate SEB, then the estimate of chemopreventive efficacy will be biased and may greatly underestimate the magnitude of the achievable risk reduction. Strategies for validating biomarkers and minimizing the degree of bias in the risk reduction estimate are discussed. Finally, problems associated with phase 2 study designs commonly used for prostate cancer chemoprevention are discussed, along with possible solutions.