An Ultra-Deep Next Generation Sequencing Method for Cfdna Detection in Nsclc Patients.

Yu Yao,Longgang Wei,Yuan,Pan Yu,Xin Wu,Zhenyu Zhang,Nan Duan,Jiaqi Zhu,Xuehong Meng,Yang Yu,Kejun Nan,Zhi Jiang
DOI: https://doi.org/10.1200/jco.2016.34.15_suppl.e23084
IF: 45.3
2016-01-01
Journal of Clinical Oncology
Abstract:e23084 Background: Targeted therapies based on gene status have dramatically improved PFS and OS in NSCLC patients. However, about 1/3 of advanced NSCLC patients can’t get enough tissue specimen to do gene sequencing. Circulating tumor DNA (ctDNA) offered a potential alternative. Since ctDNA is usually rare, novel ultrasensitive multiplex technologies are needed for ctDNA sequencing. Methods: We have developed an ultra-deep NGS method for somatic mutation detection. To validate the accuracy of SNV/Indel and fusion detection, reference samples were used, which were constructed by mixing DNA extracted from wild type human cell line with plasmid containing known SNV/ Indel, in dilution ranging from 0.5% to 10%, or mixing with DNA from cell lines carrying known fusion, in dilution ranging from 0.05% to 10%. Abundance of mutations were confirmed by 3D digital PCR. A panel including 23 cancer related core genes was used in the test, with full exon tiling, introns where rearrangement usually happen were also included. Finally, 41 sets of matched tumor and blood sample from IIIB-IV stage NSCLC patients were detected, to evaluate the concordance of test results between tumor tissue and ctDNA using the method. Results: By using the reference sample, our pipeline detected all the SNV /Indel over 1% with 100% sensitivity and 100% specificity, for 0.5%-1% mutant, the sensitivity and specificity was 90% and 90% respectively. Moreover, the abundance of mutant detected by NGS was consistent with 3D digital PCR results(R2= 0.828). All the fusions over 0.1% were detected; fusions over 0.2% were stably detected through all the 5 replicates. Using this pipeline to detect 41 sets of NSCLC patient samples, 30/41 tumor tissues had at least one mutation, 21/30 were detected in the matched blood sample, with 100% specificity, and the total consistency was 78%. For EGFR gene, 20/41 tumor tissue detected mutations, 15/20 were detected in the matched blood sample, the concordance was 88%. Seven fusions were detected, 5/7 were consistent between plasma and tumor tissue. Conclusions: We developed a reliable ultra-deep NGS method to detect ctDNA. This novel technology has a great potential for noninvasive gene mutation detection.
What problem does this paper attempt to address?