Integration of DNA- and RNA- based methods for prediction of response and survival after neoadjuvant chemotherapy for breast cancer
Final Report Abstract
The report covers a research grant at the Department of Translational Molecular Pathology of the University of Texas – MD Anderson Cancer Center, Houston, TX, USA under mentoring of Professor W.F. Symmans. The overall aim of the research project was to develop assay(s) for the prediction of response and survival after therapy for breast cancer using formalin-fixed and paraffin-embedded tissue (FFPE). The aims and methods described were adjusted according to current developments in the field of breast cancer research and the focus of the hosting research group. The main part of the report describes the development of SETER/PR, a robust 18-gene assay for the prediction of survival following endocrine therapy for hormone receptor-positive/HER2-negative (HR+/HER2-) metastatic breast cancer (MBC). We identified genes correlated with ESR1 and PGR expression in 389 HR+ breast cancer samples (U133A microarrays) and ranked them according to their pre-analytical and analytical reproducibility in datasets capturing different aspects of reproducibility. Eighteen target genes and ten reference genes were selected and summarized as the SETER/PR index. The index was then evaluated in a set of 140 biopsies from distant metastases of hormone HR+/HER2-breast cancer, and in additional pre-analytical and analytical datasets. SETER/PR was translated to a customized format for application to FFPE tissue (Affymetix Quantigene Plex). SETER/PR was theranostic for longer progression-free survival (PFS) in HR+/HER2- MBC in patients receiving endocrine therapy (HR 0.609, 95%CI 0.475 – 0.782, p = 0.003), but not in patients receiving chemotherapy. SETER/PR remained theranostic in the relevant subgroup of patients with clinical history of previous sensitivity to endocrine therapy and newly relapsed disease (HR 0.464, 95%CI 0.287 – 0.749, p = 0.002). The PFS and overall survival (OS) was longer in those with high SETER/PR (by 11 and 30 months, respectively, p < 0.0001). SETER/PR remained theranostic for in a multivariate analysis adjusted for PGR status, presence of visceral metastases, and number of previous relapse events (HR 0.578, 95%CI 0.350 – 0.956, p = 0.033). The concordance of SETER/PR index between microarray and the customized assay using FFPE tissue was 0.980 (n = 31). The SETER/PR index is a new assay to predict PFS and OS for patients with HR+/HER2- MBC who receive endocrine therapy. The assay is applicable to FFPE tissue sections from small biopsies of metastases. The results are currently validated in independent validation studies. The second part describes a study evaluating the inter-observer reproducibility of the residual cancer burden (RCB) index. RCB is a continuous index for the evaluation of response to neoadjuvant chemotherapy for breast cancer based on histopathological parameters. RCB of 95 cases treated in a randomized neoadjuvant trial was reviewed independently by 5 pathologists. The overall agreement was very good (overall concordance correlation coefficient = 0.93 (0.91 – 0.95)) and the RCB by each pathologist was prognostic for disease-free and overall survival. We concluded that RCB is a robust and meaningful clinical endpoint that is applicable in the routine pathology setting. The last part describes a systematic evaluation of quantitative and qualitative changes in gene expression according to the type of tissue (cytology vs. solid tissue) and time of sample collection (before vs. after surgery). Ultrasound-guided fine-needle aspirations and core biopsies were obtained from 22 patients at the time of diagnosis with invasive breast cancer. After surgery, additional cytology and tissue samples were collected from the resection specimens. Samples were profiled on U133A microarrays. A significant number of genes (25%) was differentially expressed when cytology and tissue samples were compared. Transcripts with a functional or spatial annotation to the extracellular matrix were enriched. Only a few transcripts were differentially expressed when pre- and post-surgery samples were compared and those were most likely associated to changes due to prolonged cold ischemic time. SETER/PR had no systematic bias in a 2-way analysis taking both conditions into account. Overall reproducibility was excellent (ICC = 0.940, κ = 0.884).
