The elucidation of multiple potential targets in cancer cells and the devel
opment of multiple target-based antineoplastic agents provide unique challe
nges in clinical trial design. Many of these agents are predicted to have c
ytostatic as opposed to cytotoxic effects and thus the traditional surrogat
e endpoint of radiologic tumor shrinkage may be inadequate. The ethical and
safety issues of obtaining multiple tumor biopsies further complicate the
assessment of appropriate target inhibition in patients. We discuss specifi
c issues that need to be addressed during preclinical, phase I, II, and III
testing of these agents. We propose clinical trial designs, including a ra
ndomized discontinuation design during phase II evaluation, that may be par
ticularly useful for cytostatic antineoplastic agents.