Abstract Introduction Chromosomal rearrangements generating gene fusions are more common in pediatric malignancies compared to adults, and possess diagnostic, prognostic and therapeutic value. Traditionally chromosomal rearrangements including structural variants (SVs) are identified using karyotyping and fluorescence in situ hybridization (FISH); however, these methods are not sensitive for small rearrangements and single nucleotide variants (SNVs). The need to detect these cryptic SV and SNVs in pediatric cancers demands the development of assays to analyze complex RNA molecules. Methods Primary bone marrow samples obtained from Nemours Biobank were used for RNA isolation. Targeted error-corrected RNA sequencing using ArcherDX HemeV2 kit was conducted on 30 primary pediatric leukemia samples and their corresponding mouse passaged xenograft samples. RNA data (fastq) was analyzed via a custom cloud environment leveraging ArcherDx Version 5.1.3 software. The gene fusion data produced by the Archer panel was initially correlated with diagnostic FISH data available for each primary sample. Results Ten out of 30 primary bone marrow samples possessed gene fusions detected by routinely tested FISH probes for diagnostic purposes, and including ETV6-RUNX1, BCR-ABL1, TCF3-PBX1, and KMT2A. In addition, this approach detected cryptic gene fusions in 10 samples that were negative for chromosomal rearrangements via FISH, including SPTAN1-ABL1, RUNX1-MKL1, NUP98-NSD1, P2RY8-CRLF2 and TCF3-HLF. The remaining 10 samples, which did not possess detectable gene fusions, showed abnormal exon usage and domain duplications for several, key oncogenes along with novel mutations. A comparison of the primary sample and mouse passaged xenograft sample revealed that majority of gene fusions representing the abundant clone remained consistent between the primary and xenograft sample in secondary and tertiary passages. Certain gene fusions representing minor clones appeared and disappeared in xenograft samples and subsequent passages in mice in comparison to the primary patient sample, highlighting the heterogeneity of the disease. Thus, the presence of major driver mutations at similar allelic frequencies in xenografts compared to primary samples and over multiple passages confirms the utility of xenograft models for preclinical drug testing. Discussion Using a novel approach that utilizes targeted error-corrected sequencing, all the aberrations detected by clinical diagnostic testing were verified, plus several novel fusion events were identified with high confidence. This method also validated the concordance between primary and xenograft samples. Characterization of these novel cryptic fusions and exonal variants in leukemogenesis will enable identification of new drug targets and prognostic factors for pediatric leukemia. Citation Format: Sonali P. Barwe, Anilkumar Gopalakrishnapillai, Nitin Mahajan, Todd Druley, E. Anders Kolb, Erin L. Crowgey. Identification of novel fusion genes and expression variants in primary and patient-derived xenograft samples of pediatric leukemia using error-corrected RNA sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2076.