New Paper on Contrived Materials in Liquid Biopsy Validation from BLOODPAC
When it comes to new technology, demonstrating that it consistently works as intended can be one of the biggest hurdles to overcome. With new diagnostic tests, this challenge is compounded by the limited availability of patient samples. A new paper by BLOODPAC members shows that contrived materials – commercially available stand-ins for patient samples – can be used to validate assays while minimizing the burden of sourcing biologic materials.
This new work, published in the Journal of Molecular Diagnostics, demonstrated that contrived materials performed comparably across study sites and technologies, increasing the field’s confidence that contrived materials can be reliably used in the assessment of liquid-biopsy workflows.
Although each company and lab does it differently, every DNA-based liquid biopsy assay is based on the detection of very small proportions of circulating tumor DNA (ctDNA) in the bloodstream. The reproducibility of a given liquid biopsy test depends on its ability to consistently detect tumor DNA at a wide range of concentrations, even when very little tumor DNA may be present. Multiple rounds of testing are needed to show an assay works as intended. But patient samples are precious, difficult to obtain, and may not reflect the necessary range of ctDNA concentrations needed for thorough performance characterization. Contrived materials can bridge the gap, providing a readily available and tunable mimic for patient blood samples that can capture much of the biological complexity and allow manufacturers to validate tests across multiple technologies and clinical laboratories.
The laboratories in this study tested contrived materials from Horizon Discovery and SeraCare Life Sciences, both as buffered cfDNA and plasma-mimetic samples. Each sample consisted of a fixed input of fragmented human cell line genomic DNA, with the variant-allele frequency (VAF)—the proportion of molecules with cancer-specific mutations— set at 0.1%, 1.0%, or 5.0% VAF per sample. Nineteen variants from the SeraCare materials and 8 variants from the Horizon materials were tested. Sensitivity—a test’s ability to correctly identify an individual with disease as positive—and specificity—a test’s ability to correctly identify an individual who does not have disease as negative—were assessed at each VAF level. Concordance between buffered DNA and plasma samples within each test site, as well as between test sites, was also determined.
In general, all participants reported comparable specificity—between 0.75 and 1.0—across samples and VAF percentages. Sensitivity was more varied, with average values between 0.8 and 1 for 1% and 5% VAF samples, and was uniformly lowest across test centers at 0.1% VAF. Concordance within a given test site and between test sites were high at VAFs of >0.1%, but reduced at 0.1% VAF. Although not surprising, given the existing literature on the difficulty of detecting low-frequency variants in circulation, the reduced performance of all tests at the lowest VAF points to the usefulness of contrived materials at capturing essential validation information across liquid biopsy assay workflows.
All data from this study is publicly available on the BLOODPAC Data Commons, to further continued collaboration and objective inter-laboratory comparisons for liquid biopsy assay development. Future work will focus on showing consistency between different tests using both biologic and contrived materials.
Read more here!
