Background: Gene fusions are major drivers of cancer and their accurate detection is key for supporting diagnosis and therapy selection. There are more than 65,000 annotated gene fusion events. Current fusions detection strategies are limited to amplicon panels querying only a small number of genes, or whole exome capture based assays, which can reduce sensitivity. With these technologies, multiple small panels are required to capture all relevant fusions, which is prohibitive for small biopsies like fine needle aspirates. Hence, clinical grade assays able to accurately detect a comprehensive set of the most clinically relevant gene fusions for solid tumors are needed.

Methods: We isolated total nucleic acid (TNA) from 141 FFPE tumor specimens or a control sample (Seraseq) and performed pair-ended, strand-specific hybridization-based RNA sequencing on a Next-Seq 500. We selected 252 fusion genes from NCCN and WHO guidelines, published clinical studies, and the 120 most frequent curated fusions in solid tumors from COSMIC (v91). These fusions are clinically relevant to most frequent cancers, including breast, colorectal, lung, lymphoma, pancreatic, prostate, salivary gland, sarcomas, and thyroid cancers. Chimeric probes were synthetized targeting fusion RNA sequences for 2230 selected breakpoints and exon junction regions of aberrant RNAs, including EGFRvIII, MET exon 14 skipping, ARv7 and ARv9. We also targeted the full coding region of 27 genes whose change in expression may suggest translocations and mutations, or have diagnostic value. Fusions were called by a custom pipeline using 3 fusion callers and a machine-learning algorithm. The PCR-based Archer FusionPlex assays and RT-PCR followed by Sanger were used as orthogonal validation methods.

Results: This single RNA fusion assay can detect at least 1194 unique known fusions pairs involving 1104 genes, as compared to 63 or 47 fusion genes from other commercially available NGS assays. On 141 FFPE samples from various tumor types, we detected 100/106 fusions reported in our CLIA laboratory Archer Sarcoma and Comprehensive Thyroid & Lung (CTL) panels. We also identified 41 additional high confidence fusions missed by Archer but confirmed by Sanger sequencing. Thus, the assay has 95.9% (141/147) sensitivity and 100% specificity, as all fusions were confirmed by either orthogonal assay. We also detected MET exon 14 skipping, EGFRvIII and ARv7 variants. Importantly, in two samples we detected MET exon 14 skipping not predicted from DNA mutation analysis showing the sensitivity of the approach. Notably, several of the 41 additional fusions are novel and were targeted by only one of the breakpoints.

Conclusion: We developed a novel and efficient breakpoint targeted fusion detection RNA-seq assay from extracted TNA from FFPE samples that can comprehensively profile thousands of clinically actionable RNA fusions and aberrant RNAs in solid tumors.

Citation Format: Fernando J. Lopez-Diaz, Steven P. Rivera, Chenyin Ou, Christophe Magnan, Brad Thomas, Tibor Gyuris, Yanglong Mou, Segun Jung, Madhuri Paul, Forrest Blocker, Shari Brown, Jacqueline Lekostaj, Ryan Bender, Sally Agersborg, Lawrence M. Weiss, Vincent Funari. A novel comprehensive breakpoint-targeted assay for clinically actionable RNA fusions and aberrant RNAs in solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2238.