Background: Liquid biopsies are increasingly utilized as a non-invasive tool in precision oncology to assess tumor mutational profiles in order to select targeted therapies, detect treatment resistance, and monitor disease progression in cancer patients. Additionally, liquid biopsies may provide a more comprehensive representation of tumor heterogeneity than standard tissue biopsies. However, limitations such as scarcity of circulating tumor DNA (ctDNA) and/or variants at low frequencies can be technically challenging to detect by next-generation sequencing (NGS) assays. Here, we use NGS to detect greater than two KRAS/NRAS mutations coexisting in single samples at low variant allele frequencies (VAFs).

Methods: The Tempus xF liquid biopsy NGS assay is designed to detect actionable oncologic targets spanning 105 genes in plasma. The assay was validated to reliably detect single-nucleotide variants at 0.25% VAF, indels and copy number variants at 0.5% VAF, and fusions at 1% VAF with 96.2%-100% specificity and 97.4%-100% sensitivity. Pre-designed digital PCR assays were modified to measure 10ng of cell-free DNA (cfDNA) on a droplet-digital PCR (ddPCR) platform.

Results: Overall, we report 100% positive predictive value and high correlation between ddPCR results and xF VAF, as well as in individual variants, such as KRAS G12D. Unexpectedly, we detected more than two coexisting KRAS/NRAS mutations at a low VAF in the plasma samples. To orthogonally confirm these results, ddPCR was deployed to independently measure the presence of each cfDNA variant with a sensitivity of 0.09% VAF. Subsequent ddPCR analysis of all targeted variants were concordant with NGS results.

Conclusions: The occurrence of multiple KRAS and NRAS mutations in a single sample is quite uncommon and may be falsely interpreted as an NGS artifact. However, verification of this phenomenon by ddPCR confirmed the validity of the NGS liquid biopsy approach. These results highlight the capability of the Tempus xF assay to detect low-frequency variants, including those that fall below the validated detection threshold, which is essential for the diagnosis of early disease.

VIEW THE PUBLICATION