Molecular-targeted agents are increasingly used for the treatment of cancer. However, the attrition rate for drugs that enter early clinical trials is higher than for other branches of internal medicine, suggesting that preclinical development has not been successful in identifying agents that can modify the outcome of human cancer. New preclinical strategies including genetically engineered mouse models and small-interfering RNAs are being used to evaluate novel agents, and have aided in the development of compounds, such as inhibitors of phosphatidylinositol 3-kinase or poly(ADP-ribose) polymerase. In addition, these techniques have helped in the identification of promising combinations of targeted drugs. In this Review, we describe methods for the preclinical evaluation of novel agents, their limitations, and strategies for improvement.
© 2011 Macmillan Publishers Limited. All rights reserved