Despite improvements in cure rates, cancer is still the leading cause of disease-related deaths among children in high-income countries. As childhood cancers are rare, concerted efforts are needed to identify relevant genomic alterations. The centralized position of the Princess Máxima Center for Pediatric Oncology enables uniform data generation from patients in The Netherlands. Whole-genome sequencing (WGS) and RNA-sequencing are routinely performed, leading to a homogeneous representative national dataset. The dataset consists of 1,539 primary tumor samples, covering 36 hematological tumor entities, 90 solid tumor entities and 36 brain tumor entities.

Likely early tumor driver events were derived from single nucleotide variants (SNVs) as well as copy number alterations (CNAs), structural variants (SVs) and gene-fusions using paired tumor-normal WGS samples. Subsequent enrichment testing of SNVs, CNAs and SV breakpoints within genes resulted in a list of genes with corresponding p-values per variant type which were then combined to harmonic mean p-values. Gene-fusions were called on RNA-seq and sub-selected for high-confidence events with WGS-supporting SV calls.

We identified 813 candidate driver genes (q-value < 0.01), with ~98% of tumor samples containing an alteration in at least one candidate driver gene. 30% of the candidate driver genes show contributions of multiple mutation types such as CNAs, SV breakpoints and SNVs, highlighting the value of integrating different mutation types for tumor driver identification. About a third of the candidate driver genes appear primary group specific. Of the candidate driver genes affected by an SV breakpoint, ~23% could be linked to a gene-fusion of which ~17% contain a known oncogene. Preliminary functional analysis further confirms cases of amplification driven overexpression of oncogenes.

This study provides the first comprehensive landscape analysis of tumor driver events in a nationally representative cohort of primary pediatric cancer patients. Our analysis reveals recurrently mutated candidate driver genes across most tumor samples, with some of those directly impacting gene expression. Over the coming weeks, we expect to finalize functional analyses as well as finish analyses aimed at identifying early driver events, thereby enabling a more comprehensive characterization of early tumorigenic drivers. This will enhance understanding of tumor biology and inform future clinical decision-making.

Additional Maxima author contributions include Ianthe van Belzen, Fleur Wallis, Anastasia Spinou, Roula Farag, John Baker-Hernandez, Alex Janse, Shashi Badloe, Marcel Santoso, Eugene Verwiel and many others.