Correlation of oncogenic mutations in circulating cell-free DNA (cfDNA) and tumor tissue through a multiplex sequencing platform in patients under consideration for phase I trials.

Abstract

6 Background: Previous studies suggest that tumoral cfDNA is a potential biomarker in cancer. Methods: Prospective collection of samples from patients (pts) referred to a phase I trials unit, in 2 cohorts: “A”(manual isolation of cfDNA, Sep09-Dec10) and “B”(robotic isolation, Jan11-Sep11) and from 41 healthy volunteers (HV). CfDNA and tumor DNA from matched formalin-fixed paraffin-embedded (FFPE) tissue were isolated, quantified and analyzed for 238 oncogenic mutations (OM) in 19 oncogenes using Sequenom OncoCarta Panel 1.0. Mutation data were used for trial allocation when available. Statistical analyses used Pearson’s Chi-squared test and Mann-Whitney test. Results: 280 pts were included: breast (60; 21%), ovarian (44; 16%), and colorectal (42; 15%). 265 (95%) plasma and 194 (69%) FFPE samples were suitable for analysis. In 181 pts (65%) both samples were available for comparison. Median turnover time for cfDNA 7 days (range 5-14). Median [cfDNA] (concentration; ng/ml) in pts was higher than in HV (“A” 21 v 6.4;p<0.01, “B” 34 v 24;p<0.01). Detection of OM in cfDNA was more frequent when [cfDNA] was above the HV range (67% v 33% p=0.002). Among 181 pts with both samples analyzed, 53 OM were detected in cfDNA, with high concordance with tumor analysis (Table), especially for RAS in colorectal tumors (10/11 91%) and PIK3CAin breast (4/4 100%). In contrast, 29 pts (16%) had OM detected in tumor but not in plasma. The rate of concordance between OM detected in plasma v FFPE was independent of [cfDNA]. 154 pts were enrolled in phase I trials and 26 were given a matched drug according to the mutational analysis. Conclusions: cfDNA may be a useful tool to detect OM in a time frame suitable for clinical decision-making. Concordance of OM found in cfDNA with FFPE samples is high; discordance in this study may be explained by limitations of multiplexed-sequencing techniques, tumor heterogeneity or biologic evolution of the disease. [Table: see text]