Adjuvant endocrine therapy reduces the risk of relapse and death from early stage hormone receptor positive breast cancer. However, tamoxifen is only partially effective because of the development of tumor resistance. Aromatase inhibitors (letrozole, anastrozole and exemestane) are also prone to the development of resistance but the pharmacologic action (estrogen deprivation) is distinct and so different mechanisms may be responsible. The problem of endocrine resistance can be directly studied in patients by examining the relationship between predictive tumor biomarkers and clinical outcome. In an example of a prospectively planned biomarker study, tumor samples were examined from a randomized trial of neoadjuvant endocrine treatment in which letrozole proved more effective than tamoxifen in terms of the rate of breast conservation and tumor regression. Interestingly letrozole was more effective at all levels of ER expression and was particularly more efficacious than tamoxifen for tumors that expressed HER1 and/or HER2 (with ER). This suggests that HER1/2 predicts primary tamoxifen resistance and relative sensitivity to potent estrogen deprivation, perhaps because HER1/2 signaling promotes the partial agonist effects of tamoxifen. A Phase 2 study of neoadjuvant letrozole is now underway to focus on gene expression profiling as a mechanism to further investigate the transcriptional programs that underlie resistance and sensitivity to estrogen deprivation. Expression profiles taken at baseline and after 1 month of therapy reveal dramatic reductions in the expression from genes responsible for DNA replication and synthesis, cell cycle progression, suppression of apoptosis and tissue invasion. When enough profiles have been generated it should be possible to detect complex interaction patterns that correctly reclassify ER+ disease into treatment responsive and resistant categories with high probability.