Classification of Breast Cancer Through the Perspective of Cell Identity Models.

The mammary epithelium has an inner luminal layer that contains estrogen receptor (ER)-positive hormone-sensing cells and ER-negative alveolar/secretory cells, and an outer basal layer that contains myoepithelial/stem cells. Most human tumours resemble either hormone-sensing cells or alveolar/secretory cells. The most widely used molecular classification, the Intrinsic classification, assigns hormone-sensing tumours to Luminal A/B and human epidermal growth factor 2-enriched (HER2E)/molecular apocrine (MA)/luminal androgen receptor (LAR)-positive classes, and alveolar/secretory tumours to the Basal-like class. Molecular classification is most useful when tumours have classic invasive carcinoma of no special type (NST) histology. It is less useful for special histological types of breast cancer, such as metaplastic breast cancer and adenoid cystic cancer, which are better described with standard pathology terms. Compared to mice, humans show a strong bias towards making tumours that resemble mammary hormone-sensing cells. This could be caused by the formation in adolescence of der(1;16), a translocation through the centromeres of chromosomes 1 and 16, which only occurs in humans and could trap the cells in the hormone-sensing state.