Biomarker subpanel customization for different use cases in acute myeloid leukemia research.

Abstract

e19026 Background: Treatment guidance and decision making according to modified RECIST guidelines often depends on characterization of these biomarkers in patients with cancer. Although panels with larger breadths of coverage are useful for initial detection of these biomarkers for evaluation of disease burden and prognosis, their cost-effectiveness decreases as treatment and disease progress; smaller subpanels may yield equally valuable information into disease progress and guide treatment in a more cost-effective manner, potentially leading to more datapoints in decision making in comparison to larger panels. Methods: QBDA is an NGS enrichment research technology that allows the enrichment of rare variants, reducing the sequencing cost while maximizing the sensitivity. Here we introduce a new capability of our platform that allows tailoring sequencing panel breath to the disease stage needs. We demonstrate the ability to generate a 384-plex NGS subpanel from a 2400-plex acute myeloid-leukemia (AML) measurable residual disease (MRD) targeting research panel. Both panels were tested on both wild-type and 0.1% variant allele frequency (VAF) AML MRD contrived samples. Results: Without any optimization of the subpanel itself, the subpanel exhibited similar or better performance to the main panel. The main panel had a ~80% uniformity of coverage whereas the subpanel had a uniformity of coverage exceeding 75%. The subpanel covers fewer genes of interest and requires fewer overall reads. Both panels had a bulk LoD of below 0.1% VAF. Conclusions: This data represents proof-of-concept for the design of small custom panels from large-scale multiplex panels for different use cases, potentially increasing cost-effectiveness.