| Literature DB >> 26970585 |
Sophie R Wang1, Simeen Malik2, Iain B Tan3, Yang Sun Chan1, Qiangze Hoi1, Jack L Ow1, Cassandra Z He1, Cindy E Ching1, Dianne Y S Poh1, Hui Maan Seah1, Katie H T Cheung1, Dharuman Perumal1, Arun G Devasia1, Lu Pan1, Shimin Ang1, Seow Eng Lee1, Rachel Ten4, Clarinda Chua4, Daniel S W Tan5, James Z Z Qu1, Yasmin M Bylstra6, Lionel Lim1, Alexander Lezhava1, Pauline C Ng1, Christopher W Wong1, Tony Lim7, Patrick Tan8.
Abstract
Targeted next-generation sequencing is becoming increasingly common as a clinical diagnostic and prognostic test for patient- and tumor-specific genetic profiles as well as to optimally select targeted therapies. Here, we describe a custom-developed, next-generation sequencing test for detecting single-nucleotide variants (SNVs) and short insertions and deletions (indels) in 93 genes related to gastrointestinal cancer from routine formalin-fixed, paraffin-embedded clinical specimens. We implemented a validation strategy, based on the College of American Pathologists requirements, using reference DNA mixtures from cell lines with known genetic variants, which model a broad range of allele frequencies. Test sensitivity achieved >99% for both SNVs and indels, with allele frequencies >10%, with high specificity (97.4% for SNVs and 93.6% for indels). We further confirmed test accuracies using primary formalin-fixed, paraffin-embedded colorectal cancer specimens characterized by alternative and conventional clinical diagnostic technologies. Robust performance was observed on the formalin-fixed, paraffin-embedded specimens: sensitivity was 97.2% and specificity was 99.2%. We also observed high intrarun and inter-run reproducibility, as well as a low cross-contamination rate. Overall assessment using cell line samples and formalin-fixed, paraffin-embedded samples showed that our custom next-generation sequencing assay has consistent detection sensitivity down to 10% variant frequency.Entities:
Mesh:
Substances:
Year: 2016 PMID: 26970585 DOI: 10.1016/j.jmoldx.2016.01.006
Source DB: PubMed Journal: J Mol Diagn ISSN: 1525-1578 Impact factor: 5.568