During the last decade, there has been increasing interest in the use of biomarkers in cancer epidemiology to enhance exposure assessment, to gain insight into disease mechanism, and to understand acquired or inherited susceptibility. To facilitate the use of biomarkers in health research, biomarkers have been divided into categories that depict the spectrum of cancer pathogenesis from exposure to disease. In this paper, we consider the epidemiological designs most suitable for the study of each type of marker. In particular, we present a two-dimensional matrix relating the biomarker categories on one axis to four different types of activities (laboratory, transitional, and etiological studies and public health applications) that develop markers and apply them in human populations. We then use the matrix to review the potential application of biomarkers in observational studies of cancer etiology, discussing the advantages, disadvantages, and logistical considerations in using biomarkers to answer research questions.