Abstract 565: Analytical performance of TruSight® Tumor 170 in the detection of gene fusions and splice variants using RNA from formalin-fixed, paraffin-embedded (FFPE) solid tumor samples

Abstract

Abstract Recent studies have highlighted the importance of gene fusions and splice variants in solid tumor profiling1. Next-generation sequencing can be an effective means of detecting these alterations in FFPE samples using RNA rather than DNA, as a single chimeric RNA transcript could result from numerous alterations in DNA2. To that end, Illumina developed TruSight® Tumor 1703, a comprehensive, hybrid capture-based NGS assay targeting 170 key cancer genes. Along with a DNA workflow, the assay includes a RNA workflow for the identification of splice variants and gene fusions. Following sequencing on the NextSeq® or HiSeq® instruments, TruSight® Tumor 170 offers an analytical pipeline which initiates variant calling. These algorithms were first optimized against the simulated read data from >350 fusions and splice variants reported in the RNA content of the gene panel. A hybrid approach of read alignment and assembly was used to enhance the fusion calling sensitivity. Deliberate filters were designed to reduce false positive calling from sequence homologs, polymerase read-through, or FFPE artifacts. For splice variant calling, a panel of FFPE non-cancerous samples were used to capture false positive mutation calls. With endogenous RNA splicing in cellular physiology, exon-boundary probes were added in the hybrid capture to enhance enrichment efficiency. To the best of our knowledge, there is not yet a standard definition for the limit of detection (LoD) in detecting gene fusions and splice variants from NGS data. We propose to define the LoD of a fusion calling and splice variant NGS panel as the lowest molecule count of a chimeric transcript that could be reliably detected with a sufficient number of supporting sequencing reads. To determine the LoD of TruSight® Tumor 170 using this definition, we mixed cell lines expressing a panel of known fusions and splice variants to measure the copy number of each chimeric transcript. Using these samples we examined the ability of the assay to confidently detect the alterations using 40 ng of RNA input. To demonstrate the analytical sensitivity and specificity of this NGS based assay, we compiled a panel of 49 mixed samples and validated the molecule count to be near the LoD of 5 copies per ng RNA input by PCR. The sensitivity was >98% for fusions and 100% for splice variants. For understanding the limit of blank (LoB) of the assay, another panel of 40 samples not harboring fusions and splice variants was also assessed by TruSight® Tumor 170. These samples demonstrated a ~97% specificity for fusion calling and >95% specificity for splice variant calling. These results indicate that the TruSight® Tumor 170 panel analysis can identify lowly expressed fusions and splice variants from a small amount of compromised RNA from solid tumor samples at high analytical sensitivity and specificity. 1 Klijn et al. (2015) 2 Maher et al. (2009) 3 For Research Use Only. Citation Format: Tingting Du, June Snedecor, Jennifer S. LoCoco, Xiao Chen, Laurel Ball, Allan Castaneda, Danny Chou, Katie Clark, Brian Crain, Anthony Daulo, Manh Do, Sarah Dumm, Yonmee Han, Mike Havern, Chia-Ling Hsieh, Tingting Jiang, Suzanne Johansen, Scott Lang, Rachel Liang, Jaime McLean, Yousef Nassiri, Austin Purdy, Jason Rostron, Jennifer Silhavy, Natasha Talago, Li Teng, Kevin Wu, Clare Zlatkov, Chen Zhao, Ali Kuraishy, Karen Gutekunst, Sohela De Rozieres, Matthew Friedenberg, Anne C. Jager, Han-Yu Chuang. Analytical performance of TruSight® Tumor 170 in the detection of gene fusions and splice variants using RNA from formalin-fixed, paraffin-embedded (FFPE) solid tumor samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 565. doi:10.1158/1538-7445.AM2017-565