Background: Alpelisib and fulvestrant are used as a combination treatment option for postmenopausal PIK3CA-mutated, hormone receptor positive (HR+), human epidermal growth factor receptor 2-negative (HER2-), advanced or metastatic breast cancer (a/mBC) patients. However, despite the presence of activating mutations in PIK3CA, the majority of patients do not derive benefit, or ultimately progress while on alpelisib therapy. Here, we investigate the genomic landscape of PIK3CA-mutated, HER2- a/mBC using next-generation sequencing (NGS) to provide insight into possible mechanisms of therapeutic resistance to alpelisib/fulvestrant and to identify potential targetable pathways.

Methods: We utilized the Tempus LENS platform to retrospectively analyze de-identified NGS data from 2,918 a/mBC patients with formalin-fixed, paraffin-embedded tumor biopsies sequenced using the Tempus|xT solid tumor assay (DNA-seq of 595-648 genes at 500x coverage; full transcriptome RNA-seq). Mutations identified included germline and/or somatic single nucleotide variants, insertions/deletions and copy number variations (gains defined as ≥8 copies). We used curated clinical data to determine HER2 and hormone receptor (ER/PR) status.

Results: Among 2,918 a/mBC patients, we identified somatic mutations in PIK3CA in 782 (26.8%). Within these tumors, 629 (80.4%) had one of the 11 mutations currently included in the alpelisib companion diagnostic, and we focused on this population (here defined as mut-PIK3CA). Of these 629, 546 (86.8%) were HER2-, with 176 (32.3%) and 370 (67.7%) derived from primary and metastatic tumors, respectively. Cases were further classified as HR+ (defined as ER+ or PR+) or triple negative (TNBC). While the majority of mutPIK3CA samples were identified in HR+ disease, 10% of the cases occurred in TNBC. Within the mutPIK3CA cohort, tumor mutational burden high (TMB-H; defined as ≥10 mutations/MB) was detected in 11.5% of samples, while microsatellite instability high (MSI-H) was detected in 0.5%. MSI-H was detected at a higher frequency in TNBC compared to HR+. Overall, the most commonly co-mutated genes among mutPIK3CA, HER2- samples were TP53 (34.6%), CDH1 (21.6%), ESR1 (12.3%), KMT2C (11%), MAP3K1 (9.5%), ARID1A (8.1%), PTEN (6.8%), GATA3 (6.6%), NF1 (5.9%), and TBX3 (5.9%) among others (Table 1); some of these genes have previously been implicated in resistance to endocrine therapy or PI3K inhibitor. In addition, in HR+ disease, metastatic samples had a higher frequency of mutations in genes implicated in endocrine resistance, such as ESR1 (18.7% vs 1.9%), ERBB2 (3.3% vs 2.6%), NF1 (6.8% vs 2.6%), compared to primary tumors. We also identified copy number gains (CNG) in several cell cycle genes, including: CCND1 (15.2%), CDK4 (2.7%), and AURKA (2.6%) (Table 1). Finally, further analyses at the transcript-level are the subject of on-going research.

Conclusions: Our study highlights that there is substantial genomic heterogeneity among mutPIK3CA, HER2- a/mBCs. Across a series of comparisons between primary and metastatic samples, as well as HR+ and TNBC subtypes, we identified a number of co-mutations that occur alongside mutPIK3CA and which could be potentially exploited by targeted therapies. Future studies are needed to assess the prognostic/predictive role of these and other candidate gene alterations.

VIEW THE POSTER

VIEW THE PUBLICATION