Technical validation of a high-sensitivity target capture NGS assay using unique molecular identifier approach.

Abstract

e13657 Background: Identification of a broad spectrum of somatic mutations is crucial to guide targeted therapy such as for non-small cell lung cancer (NSCLC) patients. In the clinical environment, it requires well validated NGS workflow both for the wet-lab and dry-lab procedures. Here we describe a high sensitivity target NGS assay to accurately capture single nucleotide variants (SNVs), short insertions and deletions (indels), copy number alterations and gene rearrangements for formalin-fixed paraffin-embedded (FFPE) NSCLC patient samples. Extensive analytical validation was performed following the checklists of College of American Pathologists. Methods: Next generation sequencing (NGS) libraries were prepared using extracted DNA from FFPE tissue NSCLC patient samples. The protocol for library generation was optimized in several steps and incorporated 10bp unique molecular identifiers (UMIs). The libraries were sequenced on Illumina HiSeq X-Ten platform. The sequence data was analyzed by an in-house bioinformatics pipeline to call somatic mutations at an average depth of 4000X. Results: We tested the accuracy of 68 clinical tumor samples that were also validated by conventional or alternative methods in the third party CAP accredited labs. We observed 100% sensitivity and 100% specificity compared with the other lab¡¯s validation results. To define the limit of detection (LOD) for different mutation types, clinical DNA samples containing different variants were diluted with normal DNA. The LODs for SNV (as in EGFR L858R) and indel (as in EGFR 19del) were 0.5% and 1%, respectively. Addressing the LOD of fusion and copy number alteration is usually challenging. Our NGS assay was able to achieve 2% LOD for gene rearrangement (fusion) and 3.5 copies for copy number amplification. The high reproducibility was also achieved by inter- and intra- replicate experiments. Our NGS assay showed better performance than other widely used commercial NGS assay panels. Conclusions: We have validated an NGS based approach with UMI technology that is able to achieve high accuracy and sensitivity as low as 0.5% for detection of somatic mutations, which will improve the clinical testing performance for NSCLC FFPE samples with low allele frequencies of driver mutations.