Development and validation of Duoseq as a novel diagnostic and companion assay for lymphoma and other cancers

3074 Background: While NGS applications of DNAseq and RNAseq have proven to be powerful tools for genomic discovery, NGS remains underutilized in the clinic. We hypothesized that the clinical translation of NGS would be greatly improved by addressing two critical issues: first, having a single assay for both DNA and RNA sequencing, and second, including all the bioinformatics to enable rapid analysis and reporting of clinically relevant findings within 2 days rather than 2-3 weeks which remains the widespread standard. We developed Duoseq to address these issues. Methods: In most cancer biopsies, RNA constitutes nearly 80% of the total nucleic acid and DNA only 20%. Duoseq incorporates interfering factors that make the ligation of sequencing barcodes to RNA much less efficient than that to DNA to, in effect, invert the proportion of DNA and RNA from the sample. This enables the efficient generation of high quality DNA and RNA libraries. The assay targets over 450 recurrently altered genes in hematologic cancers. We developed secure bioinformatics software that connects to an Illumina sequencer to render four major classes of clinical measurements: mutations, gene expression (cell of origin in DLBCL), Epstein Barr virus (EBV) status and and common translocations (BCL2, BCL6, MYC, IRF4, CCND1). Results: Technicians at three different CLIA labs were trained to perform Duoseq at the respective sites using a local Illumina sequencer. The assay was performed on 111 FFPE lymphoma cases including diffuse large B cell lymphoma, and other B and T cell lymphomas that were characterized using clinical standard assays. The standards included Sanger sequencing or Foundation One testing (mutations), in situ hybridization (EBV), immunohistochemistry or Nanostring (expression) and fluorescent in situ hybridization (translocation). The included Duoseq software provided the bioinformatics results as well as annotations. With a minimum input of 50ng of nucleic acid/sample, Duoseq generated a mean depth of coverage of 438X. We found that, compared to clinical standards, Duoseq performed accurately for mutations (95.9%), EBV status (96.5%), translocations (94.1%) and expression (100%). In addition, Duoseq provided information on clonality, specific translocation partners and identified novel fusions that were not available from standard assays. Conclusions: Duoseq accurately recapitulates the clinical workup of lymphomas replacing the need for single analyte tests such as FISH, while providing a wealth of additional data. The simplicity of the assay and the included bioinformatics software enables its performance at any clinical lab without the need for specialized personnel or computing infrastructure. We anticipate the application of Duoseq as a diagnostic assay, as well as in clinical trial selection and companion diagnostic applications owing to its high accuracy and short turnaround time.