3628 Background: Fusions and translocations account for 20% of cancer mortality globally. Maximizing their detection enhances the utility of precision medicine for various solid and hematologic cancers. Practice guidelines stress the importance of RNA sequencing. Novel assay techniques employing a comprehensive genomic profiling approach, including RNA sequencing, yield information beyond conventional DNA next generation sequencing (NGS) alone. Methods: Tumor samples (N = 1517) were assayed combining whole transcriptome (RNA) sequencing, whole exome (DNA) sequencing, and comparison of tumor sequence vs. paired normal DNA. Results were analyzed to determine the frequency of rare and common RNA fusion and variant detection. Findings were mapped to a knowledge-base of targeted treatment options. Results: Analysis detected 79 (5.2%) actionable fusions and 15 (1%) transcript variants across major solid and heme-based malignancies. Notably, we observed actionable transcript variants that are not detectable at the DNA level including: EGFRvIII, EGFRvIVa and EGFRvIVb in GBM; ARv7 in prostate, and METe14 in TNBC. Many fusion cases (42%, n = 33) had no other actionable molecular abnormalities. Novel fusions included: SLC12A/ROS1 in low-grade spindle cell neoplasm with myogenic differentiation, KANK1/NTRK2 in ganglioneuroblastoma, ETV6/NTRK3 in metastatic mammary analogue secretory carcinoma, FGFR1/SCT in germ cell tumor, ZNF33B/RET fusion in GBM, SH3BP4/ERBB4 and EML4/ALK in RCC, VTCN1/NRG1 in pancreatic cancer, and AGRN/NRG1 in cholangiocarcinoma. More common actionable fusion events included: EML4/ALK in NSCLC, KIAA1549/BRAF in pilocytic astrocytoma, FGFR2 and FGFR3 in cholangiocarcinoma and urothelial cancers and ESR1 in endocrine therapy-resistant breast cancers. The fusion events detected in heme-based malignancies included MLLT10 and MLLT4 in AML, BCR/ABL in leukemias, TCF3/PBX1 in B cell ALL, NPM1/ALK in ALCL, and novel fusion CIITA/CD274 in DLBCL. All RNA fusions and transcript variants found were matched to FDA-approved or investigational treatment options. Conclusions: Maximizing the rate of variant detection for targeted therapy relies on precise identification of common and rare fusion events. Without the addition of RNA sequencing, 15 transcript variants in our cohort would have been missed and 33 of the fusions may have gone undetected by conventional DNA NGS testing, resulting in zero targeted treatment options for this vulnerable population. Further use of comprehensive genomic profiling is vital to optimizing cancer care.
O015/#436 Comprehensive genomic profiling of low-grade serous cancers and associated clinical outcomes
