The widespread application of high-throughput sequencing methods is resulting in the identification of a rapidly growing number of novel gene fusions caused by tumour-specific chromosomal rearrangements, whose oncogenic potential remains unknown. Here we describe a strategy that builds upon recent advances in genome editing and combines ex vivo and in vivo chromosomal engineering to rapidly and effectively interrogate the oncogenic potential of genomic rearrangements identified in human brain cancers. We show that one such rearrangement, an microdeletion resulting in a fusion between Brevican (BCAN) and Neurotrophic Receptor Tyrosine Kinase 1 (NTRK1), is a potent oncogenic driver of high-grade gliomas and confers sensitivity to the experimental TRK inhibitor entrectinib. This work demonstrates that BCAN-NTRK1 is a bona fide human glioma driver and describes a general strategy to define the oncogenic potential of novel glioma-associated genomic rearrangements and to generate accurate preclinical models of this lethal human cancer.