PIK3CA Mutation Kit comparisons in FFPE breast cancer samples


Year:

Session type:

Pauline Kirk1, Amanda Davies1, James Sherwood1, Angela Dale1, Hedley Carr1, Jessica Whiteley1, Helen Ambrose1, Mark Hampson1
1AstraZeneca, Macclesfield, UK

Background

Activating mutations in PIK3CA are being studied as possible predictive markers in cancer. PIK3CA mutations occur in up to 40% of individuals with ER+ve breast cancer; these mutations occur primarily in exons 9 and 20 encoding the helical and kinase domains respectively.

Method

The performance of two PIK3CA allele specific PCR kits was assessed using DNA prepared from FFPE tissue from breast cancer patients (n=154). Mutations identified with the Qiagen PI3K mutation kit, designed to detect 4 mutations, and the Roche Molecular Systems cobasĀ® PIK3CA mutation kit, which identifies 17 mutations, were compared with data generated by Sequenom MALDI-TOF or Next Generation Sequencing (NGS) using the Illumina MiSeq Platform.

Results

Forty nine confirmed PIK3CA mutations were identified. Eight mutations were found using one method only and might represent false positives or differences in the sensitivity between methods. Three of these were confirmed by NGS. Due to the additional 13 mutations detectable with the cobas kit compared to the Qiagen kit, cobas identified 7% more verified mutation positive patients in this patient sample.

Sequenom identified one additional mutation that was not included in either kit. This was also confirmed by NGS.

Conclusion

The Qiagen and cobas kits were highly sensitive (0.95 and 0.96) and specific (0.97 and 0.99) respectively. The lower specificity of the Qiagen kit was due to the performance of the E545K/D assay that identified four mutations that were not confirmed by other methods. All of these mutations were close to the limits specified by the kit. We also observed some individuals with two PIK3CA mutations which is indicative of tumour heterogeneity.

The sensitivities of the PIK3CA commercial kits are comparable but the wider range of mutations that are detected by the cobasĀ® platform yields a higher number of mutation positive patients.