WR Miller, A Larionov, L Renshaw, TJ Anderson, JR Walker, A Krause, T Sing, DB Evans, JM Dixon
J Clin Oncol
PURPOSE: Endocrine agents, such as letrozole, are established in the treatment of hormone-dependent breast cancer. However, response rates are only 50% to 70% in the neoadjuvant setting and lower in advanced disease. Thus there is a need to identify novel markers predicting for response and to understand molecular mechanisms of resistance. PATIENTS AND METHODS: Sequential tumor biopsies were taken before and after 10 to 14 days of neoadjuvant treatment with letrozole in patients with estrogen receptor-rich breast cancer. Expression profiles on high-density microarray chips were then related to clinical responses as assessed from tumor volume measurements after 3 months of treatment. RESULTS: Of 52 patients, 37 (71%) were classified as having a clinical response to letrozole and 15 being clinically resistant. Bioinformatic analysis identified 205 covariables (69 baseline expression, 45 day 14 expression, and 91 change in expression with treatment) which differentiated between clinical responders and nonresponders. Hierarchical clustering using the variables separated responders and nonresponders into two distinct groups. Ontological assessment indicated that discriminating genes were enriched toward cellular biosynthetic processes. In particular, functional gene assessment showed ribosomal protein probes to have higher baseline expression in tumors responsive to letrozole and increased expression with treatment in nonresponding cases. CONCLUSION: To our knowledge, this is the first study to describe genetic covariables and molecular processes discriminating between tumors clinically responsive and resistant to an aromatase inhibitor. The understanding of such molecular phenotypes will be important in optimizing the clinical use of aromatase inhibitors, both in terms of identifying responsive breast cancers and developing new agents to target resistance pathways.