Population-based genomic detection of childhood cancer predisposition using newborn dried blood spots.

Targeted sequencing of 11 predisposition genes (RB1, TP53, SMARCB1, WT1, RET, SUFU, PTCH1, DICER1, APC, PHOX2B) on DNA extracted from archived newborn dried blood spots.

Within a Michigan birth cohort (1987-2020), the authors identified every child developing a solid or central nervous system malignancy by age 8 years (n = 1948), then performed targeted sequencing of 11 cancer predisposition genes on DNA from archived newborn dried blood spots. Pathogenic or likely pathogenic germline variants were found in 6.8% of cases (n = 132), most often in RB1 (n = 69) and TP53 (n = 24), followed by SMARCB1, WT1, RET, SUFU, PTCH1, DICER1, APC and PHOX2B. Approximately 1 in 27,000 newborns develops an early-onset malignancy associated with a P/LP variant. Germline variant prevalence reached 100% in medullary thyroid carcinoma, 40% in retinoblastoma and 11-30% across five further diagnoses, with strong gene-tumour specificity (p < 0.001). Such variants were rare in healthy newborn comparison datasets and in gnomAD.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The retrospective design based on archived Guthrie cards is elegant: it measures the theoretical performance of newborn screening without having to deploy it, and the 1-in-27,000 figure finally gives a discussable order of magnitude. Two caveats for the clinic: the demonstration covers children who actually developed cancer, so it says nothing about penetrance in carrier newborns who never will, nor about the overdiagnosis and parental anxiety generated at population scale. The observed gene-tumour specificity is, however, a strong argument for restricting any neonatal panel to genes where early surveillance genuinely changes outcome.

Analysis by Dr Thibaut Benquey