Abstract P4-10-15: PIK3CA mutational analysis using cell-free DNA next-generation sequencing detects activating mutations that may be missed with targeted hot-spot testing

Abstract

Abstract Background: The FDA recently approved alpelisib, in combination with fulvestrant, for HR+/HER2- PIK3CA-mutated advanced/metastatic breast cancer after trials demonstrated improved clinical outcomes with this targeted combination. The companion diagnostic, the Qiagen Therascreen, is a PCR-based kit detecting 11 single nucleotide variant (SNV) mutations limited to 3 exons in the PIK3CA gene. However, other functionally significant mutations outside these hotspots, including activating SNVs and indels, have been reported, suggesting they confer PI3K dependence and, therefore, sensitivity to PI3K inhibitors. We explored the prevalence and spectrum of PIK3CA mutations that can be identified with more comprehensive testing methods. Methods: We queried a large commercial laboratory database of clinical genomic test results from Guardant360 (Guardant Health, Inc) plasma cell-free DNA (cfDNA) next-generation sequencing (NGS) analysis of 74 genes detecting SNVs, indels, copy number amplifications, and fusions. This assay includes full exonic sequencing of PIK3CA. Clinical genomic results from patients with a diagnosis of advanced breast cancer who had at least one genomic alteration detected by Guardant360 between 11/25/2016 - 6/8/2019 were retrospectively analyzed. Results: 6940 eligible samples from 5549 unique patients with advanced breast cancer were identified; some patients had samples submitted at multiple timepoints. Excluding duplicate mutations from serial sampling, a total of 2761 nonsynonymous PIK3CA SNVs were identified in 2095 unique patients (38%); 2435 of these 2761 (88%) detected in 1982 patients (36%) were predicted to be activating. 353/1982 (18%) mutation positive patients had >1 PIK3CA activating mutation detected. Among the 2435 activating SNVs, 626 (26%) were located outside the hotspots covered by the companion diagnostic test. These 626 occurrences included over 70 unique activating mutations, with 16 unique mutations observed in >10 patients each. The most common non-hotspot mutations included E726K (117 patients), N345K (83 patients), and E453K (48 patients). Additional analysis was performed to assess for PIK3CA indels; 118 PIK3CA indels were identified in 118/5549 unique patients (2%). Predicted activating indels were identified across the PIK3CA gene, including the C2 and kinase domains and in the linking region between the adapter binding and Ras-binding domains. Conclusions: PIK3CA mutation analysis with PCR-based hotspot testing is limited to only the most common mutations and will miss as many as one-quarter of alterations that could potentially be targeted with alpelisib, an FDA-approved PI3Ka inhibitor. While additional data may be needed to determine the clinical response from targeting these alterations, molecular data and case reports suggest that these less common PIK3CA mutations are viable targets for a PI3K inhibitor. Comprehensive NGS, including plasma-based cfDNA testing, should be considered to identify the full spectrum of patients who may respond to PI3K targeted therapies. Citation Format: Lesli A Kiedrowski, Dejan Juric, Aaron I Hardin, Kristin S Price, Rebecca J Nagy, Carlos L Arteaga, Joyce O'Shaughnessy, Aditya Bardia, Massimo Cristofanilli, Richard B Lanman. PIK3CA mutational analysis using cell-free DNA next-generation sequencing detects activating mutations that may be missed with targeted hot-spot testing [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-10-15.