544PThe frequency of RAS mutation in circulating tumor DNA predicts worse survival in patients with mCRC

L. Yang,Yue Cai,Jun Zhang,Huabin Hu,Zizhao Wu,R. Deng,Yanhong Deng

DOI: https://doi.org/10.1093/annonc/mdx393.070

IF: 51.769

2017-01-01

Annals of Oncology

Abstract:Background: RAS mutations predict a worse prognosis in metastatic colorectal cancer(mCRC). However, there are few findings regarding the prognostic value of RAS mutations in circulating tumor DNA (ctDNA). We aimed to compare the concordance of genomic alterations between ctDNA and tissue biopsies and assess the prognostic value of RAS mutations in ctDNA. Methods: Gene mutational status in plasma and tissue were evaluated in mCRC patients by next-generation sequencing (NGS). Kaplan-Meier curve and Cox regression model were used to compare the progressive free survival (PFS) between different level of RAS mutations frequency. Results: of NGS testing from tumor tissue and ctDNA from 110 sequential mCRC patients were compared. Analysis of 6 gene in baseline tissue and plasma samples showed a 67.3% overall agreement. Concordance between the two platforms for KRAS, NRAS, BRAF, PIK3CA, SMAD4 and FBXW7 mutations, were 80.0%, 98.2%,97.2%,91.8%,69.1%,93.6% and 96.4%, respectively. RAS mutation rate in tissue and ctDNA were 48.1% and 29.1%. Fifty-nine patients were detected RAS mutation in tumor tissue, only thirty-six patients with plasma RAS mutation. One patient was detected ctDNA RAS mutation without mutation in tissue. Across plasma RAS gene, sensitivity and specificity were 61.0% and 99.3%, respectively. With a 48.2% cut-off rate, we divided 59 tissue RAS mutation patients into two different ctDNA RAS mutation groups(high frequency and low frequency group. Median PFS in high frequency group was 1.9 months and in low frequency group was 4.8 months(P = 0.002). In multivariate analysis considering other clinical factors(i.e. synchronous or metachronous metastasis, solitary and multiple metastases and CEA level, high ctDNA KRAS mutation frequency was independent adverse prognostic factor (HR 4.09,95% CI 1.61-10.40,p=0.003) for PFS in tissue KRAS mutation patients. Conclusions: Plasma and tissue NGS testing have a high concordance in genomic alterations. Higher rate of baseline KRAS mutation frequency predicts worse prognosis in mCRC. Both plasma and tissue NGS may be necessary to describe the complex biology of mCRC.Circulating tumor DNA testing could be a viable alternative for genotyping of mCRC and recommended for routine clinical practice. Legal entity responsible for the study: None Funding: None Disclosure: All authors have declared no conflicts of interest.