Abstract 177: RNA sequencing based gene fusion detection with oncomine comprehensive assay plus

Introduction Detection of oncogenic fusions has been of great importance for understanding tumorigenesis and for precision oncology in enhancing diagnosis and selection of targeted therapies. Herein, we describe an extended Oncomine targeted RNA sequencing assay for detection of fusion transcripts and intragenic rearrangements (exon deletion/skipping). For multiple key driver genes we also supplemented the panel with a complementary transcript-based expression imbalance assay designed to identify gene fusions in a partner agnostic manner. Methods Based on evidence from Oncomine™ Knowledgebase and collaboration with leading oncology researchers, we designed an Ion AmpliSeq™ panel to target > 1,200 fusion breakpoints in > 50 driver genes, > 40 intragenic rearrangements (e.g., MET exon 14 skipping, ARv7, EGFRvIII) in 7 genes, and 5 RNA expression controls. In addition, the panel supports detection and reporting of non-targeted fusions (i.e., novel combinations of drivers and partners). We supplemented the panel with exon tiling expression imbalance assays, using amplicons tiling the exon junctions of ALK, RET, NTRK1, NTRK2 and NTRK3 to measure 3'/5' expression imbalance signatures. We developed a bioinformatic tool to call fusions from a normalized and corrected expression imbalance profile per gene (using a baseline from normal formaldehyde fixed paraffin embedded [FFPE samples]). We optimized the gene fusion algorithms and integrated them as workflows into the Ion ReporterTM Software to facilitate the summary of the results with relevant annotations, rich data visualizations and easily interpretable reports. Results We sequenced hundreds of positive and negative fusion samples including commercial reference standards, cell lines and FFPE clinical research samples on the Ion GeneStudioTM S5 sequencer. To assess the feasibility of the combined panel, we sequenced the Seraseq® FFPE tumor fusion RNA reference, 7 fusion positive cell lines with ALK, RET, ROS1, NTRK1, FGFR1, FGFR2 and FGFR3 rearrangements, and cohorts of FFPE samples using 20ng RNA as input and successfully detected the expected fusion isoforms or other RNA rearrangements in each sample. We applied the exon tiling fusion detection method for ALK, RET and NTRK1 and observed perfect concordance between the true isoform in the positive samples and the predicted breakpoint position and magnitude of 3'/5' expression imbalance indicated by the exon tiling method. Conclusions We developed an extended, m