Abstract 542: Validation of low fraction allelic variants in plasma samples from patients with late stage colorectal cancer using droplet digital PCR: Potential clinical utility for minimal residual disease monitoring

Abstract

Introduction: Reliable detection of low mutant allele fractions (MAF) in circulating tumor DNA (ctDNA) offers clinically crucial information. Importantly, minimal residual disease (MRD) detection in solid tumors has been recognized as a powerful readout for response and early relapse prediction. Considering the high sample input requirement and assay complexity of whole exome sequencing (WES) or targeted NGS technologies, it is essential to develop a cost-effective absolute quantification approach to measure MAF with low DNA input amount. Clinically relevant gene variants present both in plasma ctDNA and tissue of origin were identified in a previous study of 25 late-stage CRC samples analyzed using the Guardant Health GuardantOMNI NGS panel (ctDNA) and WES (tumor). We further investigated the potential of using droplet digital PCR (ddPCR) to measure variants present only in ctDNA; particularly for those at low MAF which could be applied for MRD monitoring. Methods: We developed ddPCR assays for 13 variants which were detected with the GuardantOMNI panel in ctDNA and not tissue from 5 of 25 advanced stage CRC samples. MAF for the previously reported variants ranged from 64.0% (FBXW7 p.R505C) to 1.2% (TP53 R273H). Among the newly investigated variants the range was 2.2% (EPHA7 I146T) to 0.4% (PDGFRA1 pL31F and RB1 p.E440K). Other variants included KRAS G12D (30.8%), PIK3CA p.C420R (29.2%), TP53 p.R273H (1.2%) and ERBB4 p.I736T (1.1%). Positive control cDNA was designed and spiked into a pool of healthy plasma to determine the baseline wildtype levels and the sensitivity of each assay. Patient ctDNA was then measured. Results: MAF from ddPCR data and the GuardantOMNI panel showed high concordance (r2 = 0.9986) for all data points (n=13). When only considering those variants with MAF less than 5% (n=9; TP53 R273H, TSC2, EP300, PPP2R1A, LIG4, EPHA7, ERBB4, PDGFRA1, and RB1) the concordance remained high with r2 = 0.9532, thus demonstrating the accuracy and sensitivity of both platforms. All ddPCR results were achieved with at most 5ng input DNA and as low as 1ng in most cases. However, the concordance begins to weaken below 1% MAF (r2=0.4024), likely due to the limit of detection of the current ddPCR assays. Conclusions: We demonstrated that ddPCR offers a highly sensitive and purely quantitative method to measure low MAF with minimal sample input requirements, which highlights the potential use of ddPCR for MRD monitoring. As one possible MRD monitoring approach, NGS panel-based assays may identify all variants at baseline screening, followed by ddPCR as a complementary solution for MRD monitoring of single variants. Currently, there are ongoing efforts in examining longitudinal ctDNA variant changes in CRC patients receiving treatment to further confirm which variants could be used for MRD monitoring in CRC. Citation Format: Dennis O9Rourke, Danyi Wang, Juergen Scheuenpflug, Zheng Feng. Validation of low fraction allelic variants in plasma samples from patients with late stage colorectal cancer using droplet digital PCR: Potential clinical utility for minimal residual disease monitoring [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 542.