Abstract P2-08-01: Analysis of PIK3CA mutation abundance in primary breast cancer with droplet digital PCR identifies frequent sub-clonal PIK3CA mutations in ER negative and / or HER2 positive breast cancer

Abstract

Abstract Background: PIK3CA is the single most commonly mutated gene in breast cancer, with highest incidence reported in ER positive and HER2 negative breast cancer. Substantial data now suggests that breast cancers show intra-tumoural genetic heterogeneity, with apparently clonal tumours composed of multiple populations of tumour cells that, in addition to the founder genetic events common to all cells, harbour private genetic alterations. Tumours with mutations that are sub-clonal may respond less well to therapies targeting these mutations than cancers with clonal mutations. To assess how frequently PIK3CA mutations are clonal founder mutations, or may be subclonal, we assessed the abundance on PIK3CA mutation using digital PCR. Methods: DNA was extracted from frozen sections of 119 primary breast cancers, following macrodissection to achieve tumour cell content of >70%. PIK3CA mutations c.1624G>A (E542K), c.1633G>A (E545K), c.3140A>T (H1047L) and c.3140A>G (H1047R) were assessed by droplet digital PCR on a BioRad QX100 system. Exon 9 mutation assays were optimised to not amplify the PIK3CA pseudogene. Mutational abundance was calculated from the Poisson distribution, expressed as the portion of PIK3CA DNA in the sample that was mutant, and compared between breast cancer subtypes. A mutational abundance of <20% was predefined to represent low abundance mutation, that may be subclonal. Results: PIK3CA mutations were detected with abundance ranging from 80.4% to 0.0063%, with 26 cancers with an abundance >20% and 19 cancers with low abundance <20% (5 cancers with abundance 1-20%, and 14 cancer with abundance <1%). There was highly correlation between repeat experiments r2 = 0.98, p<0.0001, with 100% concordance for low abundance mutations in repeat analysis. High abundance mutations were numerically more common in ER positive HER2 negative cancers (18/65, 28%) than HER2 positive or triple negative (TN) cancers (7/54, 14% p = 0.07 Fishers exact test). Conversely, low abundance mutations were less common in ER positive HER2 negative cancers (4/65, 6%) than in HER2 positive or TN cancers (10/54, 19% p = 0.047). In cancers with a detectable PIK3CA mutation, mutational abundance was higher in ER positive cancer than ER negative cancers (p = 0.023 Mann-Whitney U test), and higher in ER positive HER2 negative cancers compared to HER2 positive or TN cancers (p = 0.0024). In ER positive HER2 negative cancers 82% (18/22) mutations were of high abundance, and likely clonal, whereas in TN or HER2 positive cancers 39% (7/18, p = 0.009) were of high abundance. Conclusion: Our data suggests that hotspot PIK3CA mutations are frequently of low abundance in HER2 positive or TN breast cancer, and may be subclonal. However, we cannot exclude the possibility that these findings represent contamination. If confirmed on an independent data set, our data suggest that identification of mutational abundance may be an important component of PIK3CA mutation assessment and the potential targeting of these mutations with PI3 kinase inhibitors. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-08-01.