Atypical hyperplastic (AH) breast lesions are currently classified and treated as benign proliferative disorders, but their presence is associated with a four-to fivefold increased risk of developing breast cancer. Currently, it is not known if an AH lesion is a marker of increased risk, or is itself a premalignant lesion. To investigate this question, we used a series of 15 microsatellite loci to analyze 15 separate AH lesions microdissected from the archived pathology specimens of subjects with no coincident or previous breast malignancy. We found that a significant subset (6/15, or 40%) of these AH lesions demonstrated evidence of monoclonal microsatellite alterations, both length variation and allele loss. These monoclonal alterations suggest that the AH lesion has already undergone genetic changes conferring a growth advantage. Thus, these AH lesions may actually be early neoplasms. We also noted that monoclonality characterized AH lesions in younger as compared with older women (44 vs. 59 yrs, P Ͻ 0.05) and that a subset of monoclonal lesions (4/6) demonstrated microsatellite alterations at more than one locus, suggesting that an undetermined type of genetic instability may play a role early in the development of abnormal breast proliferations. These findings contribute to our un-