COMP-14. TUMOR EVOLUTION DIRECTED GRAPHS IMPLY THERAPIES AGAINST MOVING TARGETS IN PAN-GLIOMAS

摘要

Precision cancer medicine aims at defining targeted treatments based on personalized mutations. While this approach has been successfully applied to a few cancers, there are several obstacles in the adoption of these approaches in complex tumors such as the gliomas. Among the main challenges that preclude the efficacy of targeted therapies are clonal heterogeneity (different cancer cells within a tumor can present diverse genetic make-up), and the dynamic nature of tumors (tumors evolve to explore niches and escape therapies). To address this challenge and to predict the evolution of gliomas, we have applied the tumor evolution direct graph framework [1] to a longitudinal cohort of genomic data from 260 glioma patients. Since cancer is a dynamic process, the optimal treatment should not only consider present characteristics of cancer cells but also the evolutionary trends, which might eventually determine the clinical outcome. To this end, our framework assembles existing computational pipelines and machine-learning approaches to learn the patterns of tumor evolution and to assign the sequential order of key driving somatic alterations in glioma progression. In addition to our previous observations [2–3], this study found that TERT promotor mutations, chromosome 1p and 19q co-deletion, PDGFA amplification, and FGFR3-TACC3 fusion are all significantly early events. Meanwhile, not limited to hypermutation and mismatch repair protein alterations, a remarkable number of cases harbor tyrosine kinase mutations (such as EGFR and MET) at the late stage, highlighting the importance of targeting moving targets in treating recurrent gliomas. More importantly, the order of somatic mutations inferred from this longitudinal cohort was also used to preclude tumor behaviors based on early alterations, which provides the possibility of early-stage intervention based on moving targets. [1] Wang, et al. eLife, 3:e02869 (2014).[2] Wang, et al. Nature Genetics, 48:768–776 (2016).[3] Lee, Wang, et al. Nature Genetics, 49:594–599 (2017).