Abstract 1379: Insights into analytical factors impacting variant concordance of plasma-tumor NGS testing

Abstract

Abstract Next generation sequencing (NGS) of circulating free DNA (cfDNA) is transforming the treatment paradigm in oncology. Sensitive non-invasive cfDNA assays could revolutionize clinical trial design with improved patient safety. However, there are substantial challenges to working with cfDNA, particularly in an early-stage setting. It has been shown that solid tumors from early stage disease shed less DNA into the bloodstream making the detection of somatic variants difficult. Because of this, the current landscape of commercial cfDNA assays are focused on advanced stage patients and may lack the sensitivity and specificity to interrogate early stage samples. To understand the limitations of the currently available cfDNA assays on early-stage oncology samples, we have designed several studies to examine the factors that impact the concordance of plasma-tumor testing and hinder the ability to detect low frequency somatic variants. Three comparison studies have been completed examining both the inter- and intra- variability of commercially available plasma-based NGS assays. Two inter-vendor studies utilized matched sets of 24 tumor-plasma samples purchased from a commercial bio-banking company. The plasmas were aliquoted into 2 ml volumes and sent to the commercial vendor for sequencing with the currently available cfDNA assay, while the tumor-normal tissues were sequenced by an external tumor sequencing vendor. The unique study design allowed for confident determination of true positive, false positive and importantly, false negative variants from the reported plasma variants. The third comparison study also used commercially purchased early-stage patient and normal samples to examine the intra-vendor assay reproducibility. Summarizing the results from all three studies, we found that the majority of discordant variants were from technical rather than biological factors. The technical factors identified include mutational bias, bioinformatic filtering, and variant annotation, while the biological factors found were clonal hematopoiesis and tumor heterogeneity. Through a detailed examination of the discordant variants from the first inter-vendor comparison, we believe that less than 15% of discordance was from biological sources. Overall these findings differ from the current literature which suggests that the sub-clonal nature of a tumor is the cause of most discordance in plasma testing. We have begun to establish guidelines to identify these technical factors in our current and past datasets, with the hope of liquid biopsy testing improvement for early-stage cancer patients. Citation Format: Daniel Stetson. Insights into analytical factors impacting variant concordance of plasma-tumor NGS testing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1379.