Current chemoprevention trial designs based on epidemiological risk assessment and occurrence of cancer as an endpoint are inefficient and expensive. Novel biomarkers are needed to facilitate the development of chemopreventive interventions. The following four categories of biomarkers may be useful in prevention trials: histologic and morphometric markers; phenotypic markers of dysregulated proliferation, differentiation, and cell loss; specific oncogenes and growth regulators which are qualitatively or quantitatively altered in breast cancers; and markers of genetic and epigenetic instability. Some of these markers will be generally useful regardless of the chemopreventive approach used, whereas others may be uniquely useful in trials of specific chemopreventive agents [e.g., upregulation of progesterone receptor (PR) expression in response to tamoxifen]. The development of these markers requires three phases of study: "Phase I": assessing the prevalence of the putative marker in malignant and premalignant tissue from individuals who have developed breast cancer; "Phase II": assessing in vivo modulation of the biomarker by the proposed chemopreventive agent; and "Phase III": applying the proposed biomarker in larger-scale trials of chemopreventive agent in high-risk populations, either before or after the development of a primary breast malignancy. The use of these biomarkers may also allow identification of novel targets for chemoprevention.