Multiomic Landscape of Gastrointestinal Stromal Tumors in a Real-World Patient Cohort of 1,427 Cases

PURPOSE: Gastrointestinal stromal tumor (GIST) is a genomically driven neoplasm with a genetic profile that determines the clinical course of the disease. However, currently available molecular data are limited because of the rarity of the disease and do not fully capture GIST clinical and biological heterogeneity. EXPERIMENTAL DESIGN: To gain deeper insights into the molecular landscape of GIST, we performed a comprehensive multiomic analysis (targeted panel, whole-exome sequencing, and whole transcriptomics) in a large real-world, multicenter cohort including 1,427 cases. Pathologic review was undertaken in KIT/PDGFRA wild-type (WT) cases. Molecular findings were correlated with clinical data and insurance claims outcomes. RESULTS: There is a complex spectrum of multilayered genetic events that converge in 3 GIST molecular subgroups: KIT-mutant, PDGFRA-mutant, and KIT/PDGFRA-WT. These alterations can be captured only by using next-generation sequencing technologies and are associated with clinical features, biological aggressiveness, and patient outcomes. Mutations in alternative genes, whether actionable or not, are seldom present and unlikely to contribute to tumor progression. By contrast, the cooperative effect of novel somatic copy-number alterations may be required for GIST evolution and progression, in addition to the core set of events involved in the current cytogenetic model of tumorigenesis. CONCLUSIONS: This molecular landscape provides a broader molecular understanding of GIST and supports a widespread use of genetic profiling for patients' clinical management.