Chromosomal aneuploidy in 25 mammographically detected breast lesions (MDBL) were determined on cytological smears using directly labeled pericentromeric probes for chromosomes 7 to 12, 17, 18 and X. The lesions included seven nonproliferative (NP) lesions, seven atypical hyperplasias (AHs), and 11 carcinomas (CAs). No other significant histological findings were identified in the remaining specimens except in two mammographically detected NP lesions, where foci of AH were present in adjacent sections; therefore, these two specimens were included in the AH lesion group (moderately increased risk lesions). Corresponding tissue sections were evaluated, and the results were correlated with fluorescent in situ hybridization (FISH) results. Monosomy was defined as the loss of one signal in > or = 15% of cells, and trisomy or tetrasomy was defined by the presence of three or more signals in > or = 3% of cells. Chromosomal aberrations were detected in 2 of 5 NP, 9 of 9 AH, and 11 of 11 CA groups. The mean number of cells with three or more signals, for all chromosomes, was 1.04 +/- 0.9 in the NP group, 8.5 +/- 9.4 in the AH group, and 20.2 +/- 5.4 in the CAs. A significant statistical difference was noted between the different groups (P = .0001). Chromosomal gain was the most common aberration and involved all chromosomes. The X chromosome was the only individual chromosome with significant differences in NP, AH, and CA groups. Chromosomal loss was observed in five specimens (20%) and involved chromosomes 8, 10, 17, and 18. The authors conclude (1) significant chromosomal aberrations can be detected in AH lesions and in NP epithelium from patients with moderately increased risk lesions; (2) numerical chromosomal aberrations tend to increase with progression of disease; (3) the frequent chromosomal gains/losses involving AH suggest that some AH may display submicroscopic features of malignancy; and (4) combined chromosomal aberrations allow for significant categorization of breast lesions, especially in cytology specimens.