Journal: Cancer discovery
This study investigates the spatiotemporal evolution of oncogene-amplified extrachromosomal DNA (ecDNA) in treatment-naive human glioblastomas (GBM) using computational modeling and tumor samples.
It reveals oncogene-specific spatial heterogeneity in ecDNA, driven by:
- Random segregation
- Variable fitness advantages
Notably, EGFR-ecDNAs tend to accumulate before clonal expansions and confer strong fitness benefits, unlike PDGFRA-ecDNAs.
Variant EGFR-ecDNAs, especially EGFRvIII, arise early from wild-type EGFR-ecDNAs and reach high abundance.
These findings highlight distinct evolutionary trajectories determined by ecDNA composition, emphasizing the need to reconsider genomic data interpretations in GBM, particularly regarding ecDNA clonality and heteroplasmy.