Childhood Cancer Predisposition and Evolutionary Constraints: Novel Lessons from Germline Genomes from 1,127 Children with Cancer

PURPOSE: Cancer predisposition syndromes (CPS) with pediatric onset are the leading known cause of childhood malignancies and are increasingly guiding clinical strategies in pediatric oncology. CPS are placed under evolutionary negative selective pressure, but pediatric pancancer studies have so far failed to investigate genomic evolutionary metrics as a guide to predict penetrance and reveal novel CPS. EXPERIMENTAL DESIGN: Germline whole-genome sequencing was performed in a 5-year prospective, registry-validated, nationwide cohort of individuals diagnosed with cancer before age 18. Evolution-guided burden analysis of private germline variants in constrained genes was compared with 125,748 gnomAD exomes. RESULTS: Across a total of 1,127 participants, 16% carried a pathogenic variant in at least one CPS gene. After genotype-phenotype matching, 9% of children in the prospective cohort (n = 651) carried a variant considered causative, a rate deemed significantly higher than in previous studies [RR, 1.54, 95% confidence interval (CI), 1.37-1.75, P = 1 × 10-14]. As predicted for a disease subject to negative Darwinian selective pressure, compared with reference adults, we found a significant excess of loss-of-function (LoF) variants in the 1,500 most constrained genes (RR, 1.54, 99% CI, 1.21-1.95, P = 4 × 10-6). Surprisingly, this excess was greater than expected, leaving a significant residual enrichment of predicted LoF variants in genes evolutionarily considered the least tolerant to damage (RR, 1.41, 99% CI, 1.09-1.80, P = 4 × 10-4). CONCLUSIONS: The high frequency of LoF variants, including in known CPS, emphasizes the need for systematic and extensive germline genomic mapping as part of the diagnostic workup of patients with childhood cancer and the linkage of such data to disease and response phenotypes to guide future pediatric oncological care and ultimately pave the way for prediagnostic interventional measures.