| Literature DB >> 29704571 |
Cleo Keppens1, John F Palma2, Partha M Das3, Sidney Scudder3, Wei Wen3, Nicola Normanno4, J Han van Krieken5, Alessandra Sacco4, Francesca Fenizia4, David Gonzalez de Castro6, Selma Hönigschnabl7, Izidor Kern8, Fernando Lopez-Rios9, Maria D Lozano10, Antonio Marchetti11, Philippe Halfon12, Ed Schuuring13, Ulrike Setinek14, Boe Sorensen15, Phillipe Taniere16, Markus Tiemann17, Hana Vosmikova18, Elisabeth M C Dequeker19.
Abstract
Molecular testing of EGFR is required to predict the response likelihood to targeted therapy in non-small cell lung cancer. Analysis of circulating tumor DNA in plasma may complement limitations of tumor tissue. This study evaluated the interlaboratory performance and reproducibility of a real-time PCR EGFR mutation test (cobas EGFR Mutation Test v2) to detect EGFR variants in plasma. Fourteen laboratories received two identical panels of 27 single-blinded plasma samples. Samples were wild type or spiked with plasmid DNA to contain seven common EGFR variants at six predefined concentrations from 50 to 5000 copies per milliliter. The circulating tumor DNA was extracted by a cell-free circulating DNA sample preparation kit (cobas cfDNA Sample Preparation Kit), followed by duplicate analysis with the real-time PCR EGFR mutation test (Roche Molecular Systems, Pleasanton, CA). Lowest sensitivities were obtained for the c.2156G>C p.(Gly719Ala) and c.2573T>G p.(Leu858Arg) variants for the lowest target copies. For all other variants, sensitivities varied between 96.3% and 100.0%. All specificities were 98.8% to 100.0%. Coefficients of variation indicated good intralaboratory and interlaboratory repeatability and reproducibility but increased for decreasing concentrations. Prediction models revealed a significant correlation for all variants between the predefined copy number and the observed semiquantitative index values, which reflect the samples' plasma mutation load. This study demonstrates an overall robust performance of the real-time PCR EGFR mutation test kit in plasma. Prediction models may be applied to estimate the plasma mutation load for diagnostic or research purposes.Entities:
Mesh:
Substances:
Year: 2018 PMID: 29704571 DOI: 10.1016/j.jmoldx.2018.03.006
Source DB: PubMed Journal: J Mol Diagn ISSN: 1525-1578 Impact factor: 5.568