LAUSR

Introduction: Determining mutations in plasma (cfDNA) is a noninvasive method of profiling tumor genomic alterations. In this study, concordance of SNVs using NGS between matched plasma and formalin-fixed, paraffin-embedded (FFPE) tissue samples was investigated across all four stages of CRC. We focused primarily on stages I, II and III to determine the feasibility of using plasma for minimal residual disease (MRD) surveillance after surgical resection and for early detection of CRC. Methods: From a CRC biobank of >9,000 subjects, 299 subjects (47 stage I, 131 stage II, 102 stage III, and 19 stage IV) were selected for whom paired treatment-naive FFPE tumor tissue and 4mL of plasma was available. MSI was determined by PCR (Promega). DNA was isolated from plasma and FFPE using cobas® extraction kits. Sequencing was performed using the AVENIO ctDNA Surveillance Kit (Research Use Only) and AVENIO FFPET Surveillance kit (under development). The AVENIO kit detects four mutation classes. In this analysis only SNVs data is presented. Three different filtering methods for concordance were used for this analysis: (i) by adaptive call, (ii) by duplex support, and (iii) by single read. Concordance is defined as at least the presence of one identical SNV between matched tissue and plasma sample. Results: The most frequently mutated COSMIC genes in the tumor tissues of this cohort are TP53, APC, KRAS, PIK3CA, and BRAF with frequencies of 59%, 53%, 20%, 19% and 17% respectively. Using the most sensitive method (“by single read”), overall concordance in this cohort was 79.26%. The concordance for various stages (Stage I to Stage IV) ranges from 50-100%. If a more stringent criterion of adaptive call is used, then the concordance ranges from 21-89%. The mean number of somatic variants for MSI-low tumors (n=247) was 5.2 compared to 15.5 in MSI-high tumors (n=52) p-value 3.3e-14. Multivariate analysis showed that in addition to the clinical stage, tumor size was the most important clinical variable associated with concordance of SNVs between the matched tissue and plasma. Conclusions: This study demonstrated an overall concordance of 79.26%. Concordance was associated with the disease stage and most significantly with tumor size and T stage. High concordance for subjects with localized disease (stages IB; IIA-IIC, and IIIA-IIIC) suggests that cfDNA sequencing can be potentially used for surveillance monitoring of patients after surgery, in particular for the detection of MRD. High tissue-plasma concordance for localized CRC may allow the use of cfDNA sequencing for early detection. Prospective clinical studies are required for validation of this application. Citation Format: Preeti Lal, John Lee, Hans-Peter Adams, Lijing Yao, Frederike Fuhlbruck1, Stephanie Yaung, Sylvie McNamara, Corinna Wostmann, Sebastian Frohler, LiTai Fang, Rainer Kube, Frank Marusch, Michael Heise, Thomas Steinmuller, Matthias Pross, Rene Mantke, John Palma, Andre Rosenthal1. Concordance of genomic single-nucleotide variations (SNV) by next-generation sequencing (NGS) in paired tumor tissue and plasma in colorectal cancer (CRC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2960.