Current status of cancer drug development: failure or limited success?

Murine leukemia models have been a major part of cancer drug screening strategies over the past 30 years. While these models have reliably identified agents active against human leukemias and lymphomas, they have been considerably less effective in identifying new agents for the treatment of common human solid tumors. This review considers the shortcomings of past approaches to cancer drug development, and describes new screening strategies designed to identify carcinoma-selective drugs.