Quantitative proteomic analysis in breast cancer

Much progress has recently been made in the genomic and transcriptional characterization of tumors. However, historically the characterization of cells at the protein level has suffered limitations in reproducibility, scalability and robustness. Recent technological advances have made it possible to accurately and reproducibly portray the global levels and active states of cellular proteins. Protein microarrays examine the native post-translational conformations of proteins including activated phosphorylated states, in a comprehensive high-throughput mode, and can map activated pathways and networks of proteins inside the cells. The reverse-phase protein microarray (RPPA) offers a unique opportunity to study signal transduction networks in small biological samples such as human biopsy material and can provide critical information for therapeutic decision-making and the monitoring of patients for targeted molecular medicine. By providing the key missing link to the story generated from genomic and gene expression characterization efforts, functional proteomics offer the promise of a comprehensive understanding of cancer. Several initial successes in breast cancer are showing that such information is clinically relevant.

Evaluation of the Predictive Performance and Regimen Specificity of a 30-Gene Predictor of Pathologic Complete Response in a Prospective Randomized Neoadjuvant Clinical Trial for Stage I-III Breast Cancer

Purpose: To prospectively evaluate in a randomized trial if a previously reported multigene predictor of pathologic complete response (pCR) to preoperative weekly paclitaxel and fluorouracil-doxorubicin-cyclophosphamide (T/FAC) chemotherapy can accurately predict pCR to neoadjuvant T/FAC chemotherapy, and if it also predicts pCR to FAC only chemotherapy. Furthermore, it is unknown if the T/FAC regimen is superior to 6 courses of FAC; therefore we compare the pCR rates for patients who receive T/FAC versus FACx6 preoperative chemotherapy.Materials and Methods: Patients with stage I-III breast cancer (n=273) were randomly assigned to receive either 12 courses of weekly paclitaxel followed by 4 courses of FAC (T/FAC, n=138), or 6 courses of FAC (FACx6, n=135) neoadjuvant chemotherapy. All patients underwent a pretreatment FNA biopsy of the tumor for gene expression profiling on oligonucleotide microarrays, and treatment response prediction (pCR versus residual disease, RD) was performed using the multigene predictor. Predicted and observed pathologic responses were compared independently in the two treatment arms.Results: The pCR rate was 19% with T/FAC and 9% with FACx6 (p<0.05). In the T/FAC arm, the positive predictive value (PPV) of the genomic predictor was 38% (95%CI:21-56%), the negative predictive value (NPV) 88% (CI:77-95%), sensitivity 63% (CI:38-84%), specificity 72% (CI:60-82%), and the AUC 0.711. In the FAC only treatment arm, the PPV was 9% (CI:1-29%), the NPV 92% (CI:83-97%), sensitivity 29% (CI:4-71%), specificity 75% (CI:64-84%), and the AUC 0.584. This suggests that the genomic predictor is regimen-specific. In a multivariate analysis including age, tumor size, nodal status, histologic grade, HER2 and estrogen receptor (ER) status and the genomic predictor, only ER status was a significant predictor of pCR.Discussion: Pathologic complete response rate was significantly higher in the T/FAC arm compared to the FACx6 arm indicating a higher efficacy of the paclitaxel containing arm. Patients who were predicted to achieve pCR to T/FAC had a significantly higher pCR rate (38%) than unselected patients (19%) or patients predicted to have RD (12%) when treated with this regimen. These results confirm that the multigene predictor can identify patients with greater than average sensitivity to T/FAC chemotherapy.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 101.