The identification and exploitation of biomarkers that may predict response to anti-cancer treatments has the capacity to revolutionize the way that patients with cancer are treated. In breast cancer, the estrogen receptor (ER) and the progesterone receptor (PgR) are known to have a significant predictive value in determining sensitivity to endocrine therapies. Tumor expression of ER or PgR is known to affect clinical outcome and this information is often used to determine a patient's optimal treatment regimen. However, the measurement of ER and PgR alone is more complex than originally thought and the impact of the recently identified isoforms of ER (ERalpha and ERbeta) and PgR (PgRA and PgRB), as well as several variant and mutant forms, upon the choice of treatment remains unclear. Therefore, ER and PgR expression alone are unlikely to determine a patient's optimal treatment regimen, particularly when the amount of 'cross-talk' between different pathways, such as the epidermal growth factor receptor pathway, is considered. In order to account for the complex cell-signaling environment that occurs in breast cancer, multifactorial techniques are needed to analyze tumor biomarker expression. The recent advances in genomic- or proteomic-based approaches has enabled molecular portraits of breast cancers to be painted, allowing biomarkers of response and prognosis to be identified and characterized more accurately than before. In the future, patients could be treated according to the molecular portrait of their tumor biomarker expression, maximizing the therapeutic benefit that each patient receives.