Head and neck cancer is a major worldwide health problem; it has been estimated that approximately 900,000 people were diagnosed with this disease in 1995. Patients are generally treated with surgery and/or radiation therapy. Treatment, especially of patients with early stage (I or II) head and neck squamous cell carcinoma, is often successful. A serious concern, however, is the fact that these patients subsequently develop second primary tumors at an annual rate of 4%-7%. Molecular analyses of premalignant and malignant tissues have produced strong evidence that clonal genetic alterations occur during the early stage of aerodigestive tract carcinogenesis. Although the roles of tobacco and diet in head and neck carcinogenesis have been the subjects of epidemiologic investigations for many years, it has only recently become possible to integrate information regarding genetic susceptibility factors into the development of comprehensive risk models for these cancers. The molecular and epidemiologic studies provide the foundation on which clinical trials can be designed to evaluate the role of retinoids and other compounds in the reversal of premalignancy and the prevention of second primary tumors (i.e., in chemoprevention). This translational approach has led to studies of the utility of intermediate end point markers, such as the nuclear retinoic acid receptors, in chemoprevention strategies. Given the rapid advances occurring in this area of research, it may soon be possible to use these biomarkers to identify patients who are most at risk for developing head and neck cancer and who are most likely to benefit from chemopreventive interventions.