The present invention is directed to the identification of predictive markers that can be used to determine whether patients with cancer are clinically responsive or non-responsive to a therapeutic regimen prior to treatment. In particular, the present invention is directed to the use of certain individual and/or combinations of predictive markers, wherein the expression of the predictive markers correlates with responsiveness or non-responsiveness to a therapeutic regimen. Thus, by examining the expression levels of individual predictive markers and/or predictive markers comprising a marker set, it is possible to determine whether a therapeutic agent, or combination of agents, will be most likely to reduce the growth rate of tumors in a clinical setting.
