Abstract PD4-3: Use of community-based next-generation sequencing (NGS) in advanced breast cancer: Identification of actionable targets to guide clinical trial selection

Background: Molecularly targeted drugs specific for mutated genes are increasingly the focus of novel clinical trials. However, molecular profiling of tumors is largely unavailable in community cancer centers, where nearly 80% of cancer patients (pts) are treated. In October 2012, Sarah Cannon Research Institute (SCRI) launched a community-based molecular profiling initiative to characterize the spectrum of molecular alterations in tumors. The profiling panel focused on potentially actionable mutations for the purpose of identifying candidates for treatment with specific targeted agents (FDA approved or investigational). Herein, we report the initial data from breast cancers (BC) profiled between October 2012 and May 2013. Methods: Metastatic breast cancer (MBC) pts > 18 years of age with ECOG PS ≤ 2 who were candidates for treatment provided consent for tumor molecular profiling. Archival tumor specimens (tissue block or 10 unstained slides) obtained from either the primary or metastatic disease were collected and interrogated by NGS (1000X average coverage) in a CLIA/CAP laboratory to detect oncogenic hotspot mutations in 35 cancer-related genes. Results were reported to the treating physician within 12 calendar days of receipt of suitable tissue and were stored in a database to enable correlation with clinical outcomes. Detection of relevant molecular abnormalities was used to identify pts appropriate for clinical trials of targeted agents. Results: As of May 31 2013, a total of 594 tumor samples were profiled, 101 (17%) of which were BC samples. 8% (8/101) of the BC samples were inadequate for assay. Of the remaining 93 samples, 60 (65%) had no mutations detected. 28% of BC had single mutations and 7% had multiple mutations. PIK3CA mutations (24%) were the most frequently identified alteration. Other genetic alterations identified included RUNX1 (4%) and FGFR3 (2%) while mutations in PIK3R1, MET, KRAS, KIT, FGFR2, HER2, BRAF, SMO, MYC, DDR2 and AKT1 were infrequent, each identified in 1 pt. Patterns and frequency of mutations in the 35 genes assayed differed in the various subtypes of BC. 6% (6/93) of BC pts with appropriate tumor mutations identified by molecular profiling have been enrolled into phase I clinical trials with PI3K or mTORC1/2 inhibitors; updated treatment results will be presented. An additional 27 patients are potentially eligible for ongoing trials at SCRI. Conclusions: This community-based molecular profiling initiative has been well accepted by patients and physicians, and provides timely results. Potentially actionable mutations were identified in 35% of BCs tested; PIK3CA mutations accounted for 70% of all actionable mutations detected. Identification of BC patients with actionable mutations may add to treatment options and improve results, and will also accelerate development of new targeted agents. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD4-3.