Extensive Clonal Branching Shapes the Evolutionary History of High-Risk Pediatric Cancers

Cancer Res. 2020 Apr 1;80(7):1512-1523. doi: 10.1158/0008-5472.CAN-19-3468. Epub 2020 Feb 10.


Darwinian evolution of tumor cells remains underexplored in childhood cancer. We here reconstruct the evolutionary histories of 56 pediatric primary tumors, including 24 neuroblastomas, 24 Wilms tumors, and 8 rhabdomyosarcomas. Whole-genome copy-number and whole-exome mutational profiling of multiple regions per tumor were performed, followed by clonal deconvolution to reconstruct a phylogenetic tree for each tumor. Overall, 88% of the tumors exhibited genetic variation among primary tumor regions. This variability typically emerged through collateral phylogenetic branching, leading to spatial variability in the distribution of more than 50% (96/173) of detected diagnostically informative genetic aberrations. Single-cell sequencing of 547 individual cancer cells from eight solid pediatric tumors confirmed branching evolution to be a fundamental underlying principle of genetic variation in all cases. Strikingly, cell-to-cell genetic diversity was almost twice as high in aggressive compared with clinically favorable tumors (median Simpson index of diversity 0.45 vs. 0.88; P = 0.029). Similarly, a comparison of multiregional sampling data from a total of 274 tumor regions showed that new phylogenetic branches emerge at a higher frequency per sample and carry a higher mutational load in high-risk than in low-risk tumors. Timelines based on spatial genetic variation showed that the mutations most influencing relapse risk occur at initiation of clonal expansion in neuroblastoma and rhabdomyosarcoma, whereas in Wilms tumor, they are late events. Thus, from an evolutionary standpoint, some high-risk childhood cancers are born bad, whereas others grow worse over time. SIGNIFICANCE: Different pediatric cancers with a high risk of relapse share a common generic pattern of extensively branching evolution of somatic mutations. GRAPHICAL ABSTRACT: http://cancerres.aacrjournals.org/content/canres/80/7/1512/F1.large.jpg.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Biomarkers, Tumor / genetics*
  • Child
  • Child, Preschool
  • Clonal Evolution*
  • DNA Copy Number Variations
  • DNA Mutational Analysis
  • Follow-Up Studies
  • Humans
  • Infant
  • Mutation Rate*
  • Neoplasm Recurrence, Local / genetics*
  • Neoplasms / genetics*
  • Neoplasms / mortality
  • Neoplasms / pathology
  • Neoplastic Stem Cells / pathology
  • Phylogeny
  • Prognosis
  • Single-Cell Analysis
  • Survival Analysis
  • Whole Exome Sequencing


  • Biomarkers, Tumor