Oncology, as a therapeutic area, is characterized by a desperate medical need for new drugs; the use of drugs that kill cells and which are consequently often toxic; and rates of failure in expensive Phase III trials that eclipse many other disease areas. The poor performance of most investigational cancer drugs implies that the standard preclinical disease models are faulty or, at least, improperly used. Some studies, however, support the view that cancer models can be highly effective, but only when selected and interpreted with care.