Adjuvant therapy decisions have become more complicated as breast cancers are further subdivided into groups that mandate different treatment approaches. Division of breast cancers into three groups, ER-positive, HER2-positive, and triple negative, while providing a general basis for treatment decisions, is insufficient to provide optimal therapy for a given patient. Multi-gene assays of the primary tumor in addition to other prognostic and predictive factors are often used to determine the molecular fingerprint of the tumor and, thus, to help guide patient therapy.

Because the rational for adjuvant systemic therapy is to eradicate distant micro-metastases that are present at diagnosis in many patients, the first step in decision-making requires an assessment of the likelihood that a given patient harbors occult distant metastases, which then translates into the risk for recurrence and death, using various validated prognostic factors. The second and equally important step is to estimate the benefit of treatment for that patient using validated predictive factors. It should be emphasized that the factors associated with risk for recurrence do not necessarily predict treatment benefit. For example, positive lymph nodes is a powerful prognostic factor and is associated with a higher risk for recurrence, but this does not necessarily mean that such tumors are responsive to chemotherapy. Considering the risk for recurrence and the estimated reduction in that risk, the third step is for the physician to have an honest discussion with the patient outlining the potential benefits relative to the side effects, toxicity, and cost of treatment, in the context of the patient’s general health and preferences, to make a final decision.

Established prognostic markers include nodal status, pathological tumor size, histologic grade, markers of proliferation, ER status, and HER2 expression. Multi-gene assays such as the 21-gene recurrence score (Oncotype DX) and Mammaprint are widely used as prognostic factors in patients with ER-positive tumors based on results of retrospective studies. Prospective studies are underway to more fully evaluate their prognostic and predictive capabilities for response to chemotherapy. Retrospective studies suggest that the 21-gene recurrence score is predictive for chemotherapy benefit in both node-negative and node-positive, ER-positive patients with high scores predicting benefit of several chemotherapy regimens. Ki-67 (MIB1) is a proliferation marker used by some as both a prognostic and predictive marker although its widespread use is limited by the lack of standardization of the assay and its interpretation. An international consortium is working to standardize this potentially very valuable marker. Molecular classification of tumors into Lumina A and B, HER2-positive, basal-like, and claudin-low is still a research tool, but additional studies to evaluate its prognostic/predictive value are underway. For the most part, these assays all differentiate between the same tumor phenotypes. Well-differentiated, hormone receptor strongly positive, HER2-negative, slowly proliferating tumors respond well to endocrine therapy but less well, if at all, to chemotherapy, whereas more poorly differentiated tumors with low or absent ER, low or absent PR, rapid proliferation, and that are HER2-positive benefit from chemotherapy but respond less well, if at all (ER-negative), to endocrine therapy. In fact, studies suggest that immunohistochemical (IHC) assays for just four gene products, ER, PR, HER2, and Ki67, may be adequate to identify these phenotypes if they are done in a very standardized manner (1). This observation needs confirmation but is potentially important in reducing the cost of care and in extending this helpful test to patients in countries with fewer resources. Data are now emerging from DNA sequencing and other genomic, epigenomic, transcriptomic, proteomic, and metabolomic studies, suggesting that in the future breast cancer may be further broken down to even more subtypes with different prognoses and responses to therapy. At present, however, these studies are at the stage of cataloguing and characterizing these alterations, and none are clinically useful for adjuvant decision-making.