Metastasis is the most devastating aspect of cancer, and the major reason for treatment failure. It is perhaps surprising, therefore, that it is only relatively recently that a wide variety of clinically relevant metastasis models have become generally available. For more than 20 years, the mainstays of cancer research were a handful of transplanted rodent tumors. A few of these (most notably B16F10 and Lewis lung carcinoma) were used as metastasis models, generally by injecting the cells intravenously to give lung colonies. The cancer research community and pharmaceutical industry could be criticized for coming up with few if any new drugs effective against solid tumor metastases. Yet is this surprising when for many years the “NCI screen” consisted of mouse ascites tumors? These are localized, “liquid” tumors, where drug access is direct and blood supply irrelevant, a poor model for solid deposits in a multiplicity of body compartments with varying vascular architecture and function. Such assays are more likely to produce drugs effective against leukemias (1 ).
