Abstract 2737: Development of a clinically actionable, ultra-sensitive multi-oncogene panel compatible with ctDNA testing in urine and plasma

Background: Circulating tumor DNA (ctDNA) is rapidly emerging as a viable alternative to tissue molecular diagnostics for the detection of actionable oncogenic mutations. In non-small cell lung cancer, identifying the emergence of EGFR T790M in patients receiving first line EGFR tyrosine-kinase inhibitors (TKIs) is a clinical need. Obtaining a tissue biopsy to determine T790M status is associated with significant cost and patient morbidity. In a retrospective analysis of the TIGER-X trial, patients had similar therapeutic response to a 3rd generation TKI, independent of the sample type positive for T790M; tissue, plasma, or urine. This indicates that non-invasive urine liquid biopsy is a viable diagnostic option (Wakelee et al, ASCO, 2016). We report on the expansion of single mutation assays to a multiplex-panel of clinically actionable mutations having analytical properties with similar ultra-sensitive detection. Methods: We previously developed two next generation sequencing methodologies with high sensitivity and specificity for ctDNA detection: 1) A series of individual ultrasensitive ctDNA hotspot assays capable of detecting ultra-short DNA fragments (31-45 base pairs) and a 0.01% lower limit of detection (LLoD). The short sized amplicons coupled with mutant allele enrichment enabled high clinical sensitivity for plasma and urine ctDNA (Reckamp et al, J Thorac Oncol., 2016) 2) A multiplex panel using sequence-specific synchronous coefficient of drag alteration (SCODA) technology that enriches for mutant ctDNA to obtain a LL0D of 0.001%-0.03%, (Pel et al, PNAS, 2009; Kidess et al, Oncotarget, 2015). By combining a multiplex of short amplicons with SCODA enrichment capabilities, we developed an assay enriched for over 200 clinically actionable driver and resistance mutations within 7 oncogenes. Results: An iterative process was used to design and optimize primer and enrichment probe length and placement to produce a panel where mean coverage across hotspot regions was within 2 to 3-fold (read uniformity) and greater than 1,000-fold enrichment of mutant DNA prior to sequencing. Technical validation was performed utilizing in-house and commercially available model systems for plasma and urine demonstrating mutation detection from inputs ranging from hundreds to fewer than 10 copies and LLoD as low as 0.001%. Clinical performance evaluation of the newly developed multiplex assay is ongoing using urine and plasma ctDNA from patients with advanced cancer. C