Studies of biomarkers of putative breast carcinogens, such as DNA adducts, have been limited by the difficulty in obtaining representative ductal epithelial cells (DECs) from breast tissue. In this feasibility study, we sought to ascertain if exfoliated DECs in breast milk could be a source of DNA for biomarker studies. Specimens (n = 38) were collected over 24 h from nursing women, and a questionnaire was administered. Cell pellets were isolated by repeated centrifugation and washing. Pellets were resuspended and incubated for 2 h, with glass adherence used to remove monocytes, resulting in an enrichment of DECs of >80%. Nonadherent cells were removed, washed, and homogenized for DNA isolation. Accurate DNA quantification was performed by 32P-postlabeling of normal nucleotides under conditions of excess ATP. Although there was wide variability in the amounts of DNA recovered, DNA yield was significantly associated with the number of weeks postpartum (P < 0.01), with optimal yield between 6 and 8 weeks after birth. There were no significant associations (P < 0.05) between the number of cells recovered and milk volume, method of collection, or the number of samples in a 24-h period per individual. This study demonstrates that breast milk can be used as a source of DECs for biomarker studies of gene-environment interaction and that sufficient DNA can be recovered to evaluate carcinogen-DNA adducts and to perform genotyping assays. Using this approach, exfoliated DECs may serve as a source of representative cells for studies of breast carcinogenesis and biomarkers of exposure, susceptibility, and effect.