Pathological Complete Response of Breast Tumors to Specific Chemotherapies Can Be Predicted

Jacquelyn K. Beals, PhD

November 14, 2007

November 14, 2007 — A phase 3 clinical trial has confirmed that gene-expression signatures can predict which of 2 chemotherapeutic regimens will most effectively produce pathological complete responses (pCRs) in breast tumors. Published early online today in The Lancet Oncology, this study is the first to evaluate genomic predictors for 2 different treatments in a phase 3 trial with a large patient cohort.

The report compared the effectiveness of a nontaxane regimen (6 cycles of fluorouracil, epirubicin, and cyclophosphamide; FEC) with a taxane regimen (3 cycles of docetaxel followed by 3 cycles of epirubicin and docetaxel; TET) in 125 patients with estrogen receptor–negative tumors. Previous studies have included mixed groups of patients with estrogen receptor–positive and estrogen receptor–negative tumors, but the former are less responsive to neoadjuvant chemotherapy.

Lead author Hervé Bonnefoi, MD, from the Institut Bergonié, University of Bordeaux, France, and chairman of the Breast Cancer Group, European Organization for Research and Treatment of Cancer, Brussels, Belgium, discussed the study with Medscape Pathology via email. He explained that the better response of estrogen receptor–negative breast tumors to neoadjuvant chemotherapy "is a clinical observation. The underlying mechanism might be the proliferation — estrogen receptor–negative tumors are more proliferative."

Coauthor David Cameron, MD, from the Anglo-Celtic Cooperative Oncology Group, Edinburgh University, Edinburgh, United Kingdom, also noted in an email to Medscape Pathology that "higher-proliferating tumors are more sensitive to chemotherapy." He added that estrogen receptor–positive cancers "express higher levels of an antiapoptotic protein, bcl-2." However, the relationship between these characteristics and chemotherapeutic response is not known.

In this study, pCR was defined as the "complete disappearance of the tumor with no more than a few scattered tumor cells detected by the pathologist in the resection specimen." However, pCRs are rare in estrogen receptor–positive tumors, and including both tumor types in a study can complicate the separation of treatment-response genes from those associated with estrogen-receptor status.

Previous studies have associated gene-expression signatures with chemotherapeutic response in cell lines and tumors. In this study, gene signatures were obtained from RNA from frozen biopsies taken at diagnosis and hybridized to microarrays. The study's goal was to confirm whether these signatures could predict the chemotherapeutic response of estrogen receptor–negative breast tumors to 2 different treatments in a large, randomized, phase 3 trial.

Patients with estrogen receptor–negative tumors were randomly assigned to the FEC (n = 66) or TET (n = 59) treatment groups. In the FEC group, 28 patients showed pCRs (42%); in the TET group, 27 showed pCRs (46%).

Clinical assignment of patients to chemotherapeutic regimens must consider the expense, toxicity, and limited population benefited by taxanes (TET treatment group). Dr. Bonnefoi summarized the adverse effects of taxane treatment as the "risk of severe infection with short ospitalization (15% – 20%), neuropathy affecting daily life (10%), [and] the price of the drug itself (approximately 10,000 Euros to be added to standard treatment)."

Response predictions were based on genes known to predict the responses of cell lines to the individual drugs. Single-drug predictions were combined into regimen-related signatures. Statistical analyses modeled the benefit of a particular treatment regimen based on the predicted probability of pCR for each regimen. Subtracting the predicted probability of pCR to TET from that of FEC, patients with values less than 0 should receive TET, and patients with values more than 0 should receive FEC.

A scatter diagram plotting the predicted probability of pCR for each patient, with TET on the horizontal axis and FEC on the vertical axis, supported the following therapeutic choices:

  • Upper right quadrant: patients will respond well to either therapy, but the negative effects of taxanes suggest the use of FEC treatment;

  • Lower left quadrant: patients will respond poorly to either therapy, but new therapeutic options should be considered;

  • Upper left quadrant: patients are predicted to respond much better to FEC than to TET; and

  • Lower right quadrant: patients are predicted to respond much better to TET than to FEC.

Overall, the hypothetical pCR based on gene signatures would be 65% to 70% compared with the 42% (FEC) and 46% (TET) rates actually observed. If this potential is confirmed prospectively, the use of gene signatures may lead to major changes in breast cancer treatment, at least for patients with estrogen receptor–negative tumors. The authors also emphasized the importance of trying new drugs in patients with a low probability of pCR to conventional treatment regimens.

Stephen R. D. Johnston, MD, PhD, FRCP, consultant medical oncologist, reader in breast cancer medicine and director of clinical research and development, Royal Marsden National Health Service Foundation Trust and Institute of Cancer Research, London, United Kingdom, praised this study in an email to Medscape Pathology. "This is a very important research article," he said, "as it starts to pull out a series of gene signatures for the first time that may be predictive for response to individual chemo regimes.

"To date," he continued, "other gene expression work has focused on overall prognosis...or prediction of response to tamoxifen — this is the first one that focuses on specific [chemotherapy] regimens based on a clear, well-conducted study in the neoadjuvant setting."

Dr. Bonnefoi, Dr. Cameron, and Dr. Johnston have disclosed no relevant financial relationships.

Lancet Oncol. Published online November 14, 2007.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.