Ductal epithelial proliferations of the breast: a biological continuum? Comparative genomic hybridization and high-molecular-weight cytokeratin expression patterns.

According to current concepts, benign proliferative breast disease (BPBD) is a direct precursor of breast cancer, in a spectrum ranging from ductal hyperplasia to overtly invasive carcinoma. In this study, comparative genomic hybridization (CGH) was used to screen ductal hyperplasia and other BPBD lesions and ductal carcinoma in situ (DCIS) for common genomic abnormalities, to test the relationship between these hyperplastic and neoplastic lesions. Immunohistochemistry for cytokeratin 5/6 was used as a diagnostic adjunct to distinguish ductal hyperplasia from DCIS. A total of 42 cases of BPBD comprising ductal hyperplasia of the usual type (n=14), papilloma (n=22), tubular adenoma (n=3), and adenosis (n=3), as well as 52 cases of DCIS, were studied. All cases of BPBD consistently displayed the presence of a subpopulation of cytokeratin 5/6-expressing basal-type cells within the proliferative lesion, whereas all of the non-high-grade and most of the high-grade DCIS lesions lacked cytokeratin 5/6-positive cells. Whereas gross genomic alterations, as determined by CGH, were undetectable in BPBD, distinct genetic changes characterized all cases of DCIS, with one exception. These results confirm the usefulness of cytokeratin 5/6 immunohistology in the diagnosis of BPBD and neoplastic breast lesions and support the view that BPBD and DCIS are not closely related entities and that BPBD is not an obligate direct precursor of DCIS.