Limitations of pharmacogenomic predictor discovery in Phase II clinical trials

Pharmacogenomic marker discovery projects are increasingly incorporated into Phase II clinical trials in the hope of identifying molecular predictors of response to therapy. At the same time, some of the inherent technical and conceptual limitations of this approach are also increasingly recognized. This manuscript will review some of these limitations. We suggest that it may be more productive to use pharmacogenomic, or other molecular data, from small Phase II clinical trials to assess the clinical utility of previously defined putative markers rather than use the data primarily for discovery. We also describe a simple clinical trial design that could accomplish this. The content of this article is based on a presentation during a symposium on 'Novel Trial Designs for Early Drug Development' at the Annual Meeting of the American Association of Cancer Research in Los Angeles, CA, USA, on April 17, 2007.