Abstract P5-06-28: Optimization and validation of PIK3CA mutation detection with droplet digital PCR in liquid biopsies of patients with metastatic breast cancer

Background Approximately 40% of oestrogen receptor positive (ER+) metastatic breast cancer harbour PIK3CA hotspot mutations, leading to an over-activated PI3K pathway. The PI3Kα-selective inhibitor, alpelisib, is FDA approved and currently available in early access program for patients with ER+ HER2-negative metastatic disease, if an activating PIK3CA mutation is confirmed. Mutational analysis on liquid biopsies may increase the number of eligible patients given the challenges of tissue procurement in metastatic setting. We optimized and validated ddPCR assays for PIK3CA hotspot mutations in cell-free DNA (cfDNA) of metastatic breast cancer patients. Patients and Methods We prospectively collected two blood samples (cfDNA Collection tubes, Roche Diagnostics) of 20 metastatic breast cancer patients at progression (N=12) or during therapy (N=8). PIK3CA mutation status was previously demonstrated by Next Generation Sequencing (NGS) performed in 4 patients on primary tumor tissue and in 16 patients on metastatic tumor tissue; 8 patients tested positive and 12 negative. Tumors were ER+ HER2-negative, triple-negative and HER2-positive in 14, 5 and one patients, respectively. After plasma-isolation, cfDNA was manually extracted with Cobas® cfDNA Sample Preparation Kit (2mL plasma) and semi-automatically with Maxwell® RSC ccfDNA Plasma Kit (2mL and 4mL plasma). Inter-run variability, intra-run variability, precision and robustness of the assays, and concordance between NGS and ddPCR for the detection of PIK3CA hotspot mutations on tumor tissue and in plasma, respectively, were assessed. Results All 20 samples were successfully processed with ddPCR. The highest cfDNA yield was obtained by Maxwell 4mL extraction method (median: 0.483 ng/µL; range: 0.140 - 10.500 ng/µL). The per-mutation sensitivity and specificity was 87.5% and 95.8%, respectively. Two samples showed discordant results between PIK3CA mutations detected by NGS on tumor tissue and by ddPCR in plasma. One sample tested positive for p.(H1047R) on tumor tissue, and for p.(E542K), p.(E545K) and p.(H1047R) in cfDNA. The other switched between p.(E545K) mutation on tumor tissue and p.(E542K) in cfDNA. Tumor tissue of samples for which the results of the NGS test were discordant with the ddPCR test on plasma, was tested again using ddPCR on the tissue. This still showed discordances in PIK3CA mutations which can be explained by tumor heterogeneity and the lower detection limit of ddPCR versus NGS. Conclusion Detection of PIK3CA hotspot mutations with ddPCR in cfDNA is feasible. We observed good concordance between NGS and ddPCR for the detection of PIK3CA hotspot mutations on tumor tissue and in plasma, respectively. In 10% of cases, discordant PIK3CA mutation status in tissue versus plasma was detected. Further investigation of different metastatic lesions in these cases is ongoing. Citation Format: Laurence Slembrouck, Demi Renders, Sara Vander Borght, Patrick Neven, Giuseppe Floris, Lien Spans, Hans Wildiers, Kevin Punie, Ann Smeets, Ines Nevelsteen, Ignace Vergote, Adriaan Vanderstichele, Isabelle Vanden Bempt. Optimization and validation of PIK3CA mutation detection with droplet digital PCR in liquid biopsies of patients with metastatic breast cancer [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 P5-06-28.