Portal de Periódicos da CAPES

NGS for precision medicine in non-small cell lung cancer: Challenges and opportunities

2016; Elsevier BV; Volume: 27; Linguagem: Inglês

10.1093/annonc/mdw363.67

1569-8041

Vladislav Mileyko, Mikhail Ivanov, Novikova Eg, Е Н Телышева, Polina Chernenko, В. В. Бредер, К. К. Лактионов, Ancha Baranova,

RNA and protein synthesis mechanisms

Recent advances in target therapy development and precision oncology researches has led to the spreading of the use of the molecular testing. The need on broad tumor profiling has been met by the NGS, though several limitations remains unsolved. FFPE tissue sections were obtained from 25 patients with NSCLC (stages III-IV). DNA libraries were prepared with the Truseq Cancer Panel (Illumina). Bowtie-2 with the following Varscan2, Strelka and Scalpel accompanied with the in-house software as well as Somatic Variant Caller (Illumina) were used for data analysis. Clinicaly-actionable mutations were identified in 13 patients (52%). Of them 8 has activating EGFR mutations. Rare EGFR exon 19 insertion was identified in one patient, which may be associated with EGFR TKI sensitivity. This mutation was not detcted employing the default software (Illumina Somatic Variant Caller) due to misalignment near the end of the reads and was successfully identified with custom pipeline. Another patient with EGFR G719 mutation harbored frameshift mutation in the 717th codon, which would eliminate EGFR activation by G719 mutation. In this case detection of signle G719 mutation would falsely indicate at EGFR TKI sensitivity. Despite the absense of matched normal tissues we were able to detect CNV employing bootstrapping, allowing to detect EGFR amplifications in two patients. Low prevalent mutations were enriched with the C/T and G/A changes which are known to be FFPE arefacts. Therefore, mutations with allele frequency lower 10% were not detectable in 13 samples (52%). In four patients low library concentration led to the increased count of high prevalent mutations. This resulted in false positive mutations including AKT1 E17K and CTNNB1 S45F, suggesting that simple mutant allele frequency cutoff can not be used to sort out FFPE artefacts. NGS allows to detect rare mutations associated with TKI sensitivity which often remain unseen using gold standard methods. Obaining low-level information it allows to exclude false positive and false negative results. Though tbioinformatic pipelines remain the major sensitivity limitating stage. We were able to perform thorough configuration complexed with internal devepments to overcome these obstacles.