Development of difluoromethyl-ornithine and Bowman-Birk inhibitor as chemopreventive agents by assessment of relevant biomarker modulation: some lessons learned.

A major goal in the development of chemopreventive agents has been to develop markers that reflect the underlying process of carcinogenesis and which are modulatable by the agent under study. An important application of such markers will be to select cohorts that are at elevated risk for cancer development, which should allow use of smaller sample sizes in definitive phase III trials as well as shorter duration (and lower cost), without loss of statistical power. Susceptibility and surrogate end-point biomarkers are particularly important in this respect. Intermediate markers are probably best assessed in terms of proportionate rather than relative risk. The systematic development of difluoromethylornithine for use in chemoprevention against human cancer has involved pilot, phase IIa and IIb trials using participants with prior colonic polyps as the study group. A unique feature of the phase IIa study was the use of a dose de-escalation design which allowed selection of the lowest effective non-toxic dose of difluoromethylornithine. The phase IIb trial now in progress is using a combination of sulindac with difluoromethylornithine; the rationale for selection of markers for this study and for a randomized phase III registration trial is discussed. We also review the findings in phase I and IIa trials of Bowman-Birk inhibitor concentrate, in which patients with measurable oral leukoplakia are the study group.