Author: Gianni L, Zambetti, K., et al.
Source: Journal of Clinical Oncology 23;29, 2005.
Introduction
Gene array and RT-PCR technology allows for the rapid evaluation of hundreds of individual genes that may play a role in predicting response to chemotherapy. Previously, such techniques were limited by the fact that fresh tissue was required for analysis, but recent studies have demonstrated that a sufficient amount of RNA can be extracted from paraffin-embedded samples.
Recently, Paik et al. (NEJM 251;27,2004) reported on the results of the Oncotype DX assay (Genomic Health). In that study of 250 potential candidate genes, the authors found 21 genes to be related to risk of recurrence in breast cancer patients. The study was designed to validate a previously established 21-gene RT-PCR assay. This assay is used as part of a recurrence score (RS) algorithm for node-negative, estrogen-receptor positive breast cancer patients who had been treated with Tamoxifen in the NSABP trial B-14. The results demonstrated that patients could be categorized as having a low, intermediate, or high risk for recurrence based on the RS resulting from this assay. The authors suggested that the RS can then be used to better select patients to receive chemotherapy, i.e.: to give chemotherapy to patients who have a sufficiently high risk of recurrence.
The present study examines 384 genes to determine whether any are associated with improved response to neoadjuvant chemotherapy in patients with locally advanced breast cancer. The authors sought to determine whether a specific group of genes correlate with response to neoadjuvant doxorubicin and paclitaxel. These genes were then assessed in a separate cohort of patients to determine their response-discriminating value.
Methods of Milan Cohort
- 384 candidate genes were selected by surveying the breast cancer literature for evidence of gene linkage to pathological processes (proliferation, invasion, apoptosis, metastasis, angiogenesis, immune surveillance, tumor suppression activity, oncogene activity, and differentiation status) or to pathways involved in chemotherapy response (metabolism, drug resistance, transporters, DNA repair). The 21 genes used in the Breast Cancer Oncotype DX RS assay were included as well.
- Women with locally advanced breast cancer who were enrolled in a study of primary chemotherapy in Milan from 1998-2002 were eligible for the study. Patients were treated prior to surgery with doxorubicin (60mg/m2) and paclitaxel (200mg/m2) every 3 weeks x 3, followed by weekly paclitaxel (80mg/m2) x 12. Following surgery, the patients received adjuvant CMF (cyclophosphamide, methotrexate, and fluorouracil), locoregional irradiation, and hormonal therapy according to standard criteria.
- 89 of the Milan patients’ paraffin-embedded samples were analyzable. After paraffin removal, RNA was extracted and quantitative gene expression was determined by RT-PCR for each of the 384 + 21 genes.
- At surgery, the primary endpoint of pathologic complete response (pCR) was determined and defined as the absence of invasive cancer in the breast (residual DCIS was allowed). All pCRs were then correlated to gene expression.
Results of Milan Cohort
- Overall, 11 patients (12%) had a pCR. Seven (23%) of 31 patients with ER negative breast cancer by IHC had a pCR, and four (8%) of the 52 patients with ER-positive breast cancer had a pCR.
- Concordance between IHC and RT-PCR measurements of ER and PR expression was high.
- Univariate analysis of the odds ratio for pathologic response for each candidate gene demonstrated 30 genes with p <0.01 and a total of 86 genes with p <0.05.
- Three prominent groups of co-expressed genes were found on cluster analysis: an ER-related cluster, a proliferation-related gene cluster, and an immune-related gene cluster. Of 86 “hits”, 46 were in one of these clusters.
- Quantitative expression of the ER gene by RT-PCR was negatively correlated with the likelihood of a pCR.
- When tested against the 21 Oncotype gene panel, increasing RS was associated with increasing pCR
Methods of MDACC Cohort
- To assess the response-discriminating value of the pCR-associated genes in the Milan samples, DNA microarray results from 82 patients with locally advanced breast cancer treated at MD Anderson Cancer Center were studied. Samples were obtained via FNA before starting preoperative chemotherapy with 12 weeks of paclitaxel followed by fluorouracil, doxorubicin, and cyclophosphamide (FAC) x 4 courses. The definition of pCR at surgery was the same.
- Microarray technology was used to analyze these samples, and cDNA was hybridized to an Affymetrix U133A GeneChip array. They used 79 of the 86 predictive genes in 179 individual probe sets.
Results of MDACC Cohort
- In the MD Anderson cohort, 21 patients (26%) had a pCR.
- Thirty-seven probe sets corresponding to 24 genes demonstrated strongly significant correlation with pCR (p<0.05); an additional 32 genes correlated with p<0.1.
- Two clusters were revealed, one with 39 cases and 5pCRs (12%) and another with 43 cases and 16 pCRs (37%).
Discussion
In the last few years, several papers have been published seeking a specific gene set that will predict for outcome or response to therapy in cancer patients. The Oncotype DX assay was tested in over 400 patients from three different populations, and further validated in the study referenced above. The current study is hypothesis- generating, but it is still too early to tell whether a subset of the 86 original genes is predictive for pathologic response. The original model was tested in 89 patients, of whom only 12% had a pCR. The authors state that with this low number, they would expect a 22% false discovery rate when testing 384 genes. It is encouraging that many of the genes also predicted for pCR in the MDACC cohort, and that the direction of effect was the same in all cases. The concept of choosing specific chemotherapy, or whether or not to give chemotherapy, based on certain genetic profiles is very interesting, and we will await further validation of this study in additional patient populations to determine its clinical applicability.
About OncoLink Contact OncoLink Privacy statement Disclaimer Link to OncoLink Home