Clinical Validation of Local Versus Commercial Genomic Testing in Cancer: A Comparison of Tissue and Plasma Concordance.

Abstract

Genomic sequencing of tumours improves patient outcomes through implementation of precision oncology. At present, genomic testing is mainly confined to research settings, with samples sent to biopharmaceutical companies for analysis. The ever-expanding catalogue approved of targeted therapies has created an urgent unmet need for local genomic testing facilities, to enable upscaling of testing. Here, we compare the outcomes of local (IonTorrent^™) and commercial (Foundation Medicine) genomic testing collected from 30 cancer patients in from plasma and tissue samples. Overall concordance was high in both tissue (98%) and plasma (94.2%). Variants identified by both platforms had a strong correlation in variant allele frequencies (VAF%): plasma: r = 0.99 p < 0.0001, tissue: r = 0.91 p < 0.0001. However, numerous low VAF% variants resulted in low positive percentage agreement (tissue 78.8% plasma 16.1%) and positive predictive values (tissue 56.3% plasma 71.4%). Local sequencing demonstrated higher fidelity in detecting fusions but low fidelity in detecting indels. Overall, this study supports the use of local genomic testing for routine molecular diagnostics but highlights outstanding issues before widespread implementation. Processing of variants detected at low VAF% and the limit of detection of assays needs to be addressed. Construction of gene panels requires careful consideration, including incorporation of markers of genomic instability.