Gene expression profiles predict pathologic complete response to preoperative chemotherapy with gemcitabine, epirubicin, and docetaxel in primary breast cancer.
Reviewer: Christopher Dolinsky, MD
University of Pennsylvania School of Medicine
Last Modified: May 14, 2005
Presenter: A. Schneeweiss
Presenter's Affiliation: German Cancer Research Center, Heidelberg, Germany
Type of Session: Scientific
- One of the benefits of using preoperative chemotherapy for primary breast cancer is that one can reliably assess the efficacy of the preoperative regimen by examining the tumor response in the surgical specimen.
- A pathologic complete response (PCR) is defined as no evidence of viable, invasive tumor cells left in surgical specimen.
- Considerable interest has developed in finding methods to predict which patients will have a pathologic complete response to preoperative therapy.
- One potential method to predict which patients tumors will respond to a particular chemotherapeutic regimen is to examine a sample of their tumor for specific patterns of gene expression.
- This study was designed to look for a gene expression profile that predicts pathologic complete response to preoperative treatment of breast cancer with a three drug regimen of gemcitabine, epirubicin, and docetaxel.
Materials and Methods
- 98 patients with T2-4 N0-2 M0 breast cancer at the
- Patients ranged in age from 18-65 years old, and had no prior therapy for their breast cancer.
- The patients were split into 2 groups, a training set (50 patients) and a validation set (48 patients).
- The training set was used to develop the particular gene expression profile in the laboratory, and the validation set was used to confirm its efficacy in predicting pathologic complete response.
- The training set received 5 cycles of gemcitabine/epirubicin, followed by 4 cycles of docetaxel given every 3 weeks.
- The validation set received 6 cycles of gemcitabine/epirubicin/docetaxel given every 3 weeks.
- The 2 groups were comparable in terms of a variety of factors including: age, tumor size, nodal status, grade, hormone receptor status, and Her-2-neu overexpression.
- Each patient underwent a pretreatment core needle biopsy to collect frozen tissue for microarray analysis.
- Small amounts of isolated breast tumor RNA were amplified in the laboratory, and expression profiles were generated using DNA microarrays carrying 21,329 unique gene specific probes.
- Using the training set, a particular combination of 512 genes in the expression profile was determined to be effective in predicting pathologic complete response.
- This set of genes included prominent, well studied genes such as BRCA1 and Ras.
- Statistical analysis on the validation set demonstrated that for predicting pathologic complete response, this particular gene expression profile had a sensitivity of 78%, a specificity of 90%, a positive predictive value of 64% and an accuracy of 88%.
- When other factors previously used to predict likelihood of pathologic complete response were analyzed using logistic regression, only one, Her-2-neu overexpression (3+), was also predictive.
- Transcriptional profiling can identify a gene expression profile in primary breast cancer that may serve as a clinically useful predictor of pathologic complete response to preoperative chemotherapy containing gemcitabine, epirubicin, and docetaxel.
- When integrated with currently known predictive factors like Her-2-neu overexpression, gene expression profiles may help predict which patients are likely to benefit from a particular chemotherapeutic regimen.
This abstract presents some exciting data regarding the utility of gene expression profiles for predicting outcomes in primary breast cancer. The authors demonstrated quite elegantly that DNA microarray technology can be used effectively for this line of research. One short-coming of this presentation was that the patients in the different sets for training and validation did not receive exactly the same chemotherapy schedules, which could impact on the results. As new gene expression profiles for other chemotherapeutic regimens are developed, it may eventually be possible to direct patients towards a tailor made regimen for their particular tumor. Even patients who receive systemic therapy after surgical resection may potentially benefit from knowledge that their tumors are likely to respond to one versus another chemotherapeutic regimen. The endpoint that was studied in this abstract, pathologic complete response, is an attractive endpoint from a researcher's perspective because it is easily examined after a short period of therapy. The final test of the efficacy of gene expression profiles will be patient related outcomes: rates of local control, metastatic control and overall survival. Time will eventually weigh in on these outcomes, but in the meantime, this line of research remains promising and should certainly be continued.