Reporter gene technology is widely used to measure activity of hormone analogs, and bioluminescent in vitro assays have allowed rapid screening of numerous chemicals either to identify new agonists or antagonists of hormones or to detect the presence of endocrine disrupters in the environment. Stable bioluminescent cell lines have been established and they provide reproducible dose-response curves and accurate determination of in vitro efficiencies of various chemicals. In vivo, however, these molecules can be metabolized, bound by proteins, or stored in fats and thus could display efficiencies different from those observed in vitro. In vivo assays, such as the uterotrophic bioassay, require numerous sacrificed animals, and responses not only are dependent on an estrogenic action but also imply other factors. For a faster assay and to avoid the use of numerous animals, we developed an in vivo biosensor constituted of stable bioluminescent cells implanted in nude mice. MCF-7 bioluminescent cell lines were chosen since their proliferation is low in the absence of estrogen and the xenograft size can thus be stable for several weeks. Luciferase gene expression was monitored noninvasively with a cooled charge-coupled device camera. Quantitative analysis allowed us to compare in vitro and in vivo actions of different estrogenic compounds (estradiol, estrone) and endocrine disruptors (ethynylestradiol, genistein, octylphenol, and 2,4'-dichlorodiphenyldichloroethylene) in the same cell lines and to follow hormone action on a living animal as a function of time. Different administration protocols have been used and good correlation was observed for most products. However, we found that ethynylestradiol was the most efficient chemical when orally administered.