Direct comparison of circulating tumor DNA sequencing assays with targeted large gene panels

Next generation sequencing (NGS) assays with large targeted gene panels can comprehensively profile cancer somatic mutations in a tumor sample. Given the rapid adoption of such assays for circulating tumor DNA (ctDNA) analysis in clinical oncology, it is essential for the community to understand their analytical performance in liquid biopsy settings. Here, we directly compared five ctDNA NGS assays, most of which having a panel of 400 or more genes, with simulated samples harboring mutations relevant to solid tumors or myeloid malignancy. Our results indicate that the detection sensitivity and reproducibility of all five assays was 90% or higher when the mutations were at 0.5% or 1.0% allele frequency, and with optimal DNA input of 30 ng or 50 ng per vendor's protocol. The performances decreased and varied dramatically, when mutations were at a 0.1% allele frequency and/or when a lower genomic input of 10 ng DNA was used. Interestingly, one of the assays repeatedly showed higher rate of false positivity than the others across two different sample sets. Multiple intrinsic technical factors pertaining to the NGS assays were further investigated. Notable differences among the assays were seen for depth of coverage and background noise, which profoundly impacted assay performance. The results derived from this study are highly informative and provide a framework to assess and select suitable assays for specific application in cancer monitoring and potential clinical use.