[Research] Can Newborn Screening Identify Childhood Cancer Predisposition?

Genomic newborn screening has, until now, been discussed largely in the context of inborn errors of metabolism and immunodeficiency disorders. A recent population-based study published in Nature Communications (Diller et al., 2026), however, examined for the first time—and quantitatively, at a population level—whether its scope could be extended to cancer predisposition. The central question is a clear one: using only DNA from the dried blood spots (DBS) collected and archived at birth, can the genetic predisposition to early-onset childhood cancer be identified at the newborn stage?

Study Design

By linking a Michigan birth cohort (1987–2020) with cancer registry data, the investigators identified 1,948 children who developed a solid or central nervous system (CNS) malignancy before age 8. DNA extracted from their archived newborn dried blood spots then underwent targeted next-generation sequencing (targeted NGS) of 11 autosomal dominant cancer predisposition genes (RB1, RET, TP53, SMARCB1, SUFU, PTCH1, WT1, DICER1, APC, ALK, PHOX2B). From a single 3.2 mm blood spot punch, 1,943 of 1,948 samples (99.7%) met quality standards for analysis.

One point worth noting is that this was a phenotype-first study: it worked backward from children who had already developed cancer to identify predisposition variants. In other words, rather than a prospective study that screened first and then tracked outcomes, it estimated the potential yield of screening at a population level.

Key Findings

A pathogenic or likely-pathogenic (P/LP) germline variant in one of the 11 genes was detected in 132 children (6.8%). Detection was most frequent in RB1 (69), followed by TP53 (24), SMARCB1 (8), and WT1 (7). No P/LP variants were detected in ALK. Notably, 130 of the 132 carriers presented with a tumor type well established as associated with the affected gene, reflecting a high degree of gene–tumor specificity (p < 0.001).

Detection rates by diagnosis showed marked specificity:

• Medullary thyroid carcinoma: 100% (6/6, all with RET variants)
• Retinoblastoma: 40% (68/168); among bilateral cases specifically, 80% (41/51)
• Choroid plexus carcinoma: 30%
• Adrenocortical carcinoma: 20%
• Pineoblastoma: 17%
• Lung/pleural malignancy: 11%
• Medulloblastoma: 11%

Based on the approximately 3.55 million births in Michigan contributing to the cohort, an estimated 1 in 27,000 newborns went on to develop an early-onset solid or brain malignancy before age 8 while carrying a variant in a predisposition gene. This prevalence is by no means low compared with conditions already included in newborn screening—for example, Pompe disease (~1/18,000), severe combined immunodeficiency (~1/59,000), and maple syrup urine disease (~1/200,000).

Carriers of predisposition variants were also diagnosed with cancer at an earlier age than non-carriers (median 14 months vs. 32 months; p < 0.001). The authors reported that roughly 80% of bilateral retinoblastoma cases could have been identified at birth.

What It Means Clinically

When predisposition is recognized at birth, syndrome-specific surveillance guidelines can be applied early. Examples include regular ophthalmologic examinations for RB1 carriers, abdominal and renal ultrasound for WT1– and DICER1-related conditions, and whole-body MRI together with ultrasound surveillance for TP53 (Li-Fraumeni syndrome). Such early surveillance has been discussed as a potential route to stage shift, reduced treatment intensity and radiation exposure, and improved survival.

That Said, Knowing Early Is Not Always a Benefit

This is a point the study itself treats with care, and it warrants faithful reporting.

First, the actual clinical benefit of surveillance varies by syndrome. For some conditions early detection translates into a clear survival advantage, while for others the evidence remains limited. Indeed, this study excluded leukemia predisposition genes, noting that the benefit of presymptomatic detection is less well established than it is for solid and brain malignancies.

Second, given the limitations of the phenotype-first design noted above, these data do not prospectively demonstrate that screening improves actual clinical outcomes. Moreover, because autosomal recessive predisposition syndromes and leukemia predisposition genes were excluded, the overall size of the genetically at-risk population may in fact have been underestimated.

Third, early identification carries potential harms as well: psychological burden on families, the burden of long-term surveillance in young children, and the possibility of overdiagnosis and unnecessary intervention arising from indeterminate findings. In addition, variant frequencies derived from a specific population (the Michigan cohort) should be applied only with caution to populations of differing ancestry and demographic structure.

How This Relates to 3billion’s Newborn Screening

Many of the 11 genes targeted in this study fall within the cancer-related genes assessed by 3billion’s genomic newborn screening test (3B-NEO). Specifically, RB1, RET, TP53, WT1, SUFU, PTCH1, DICER1, and APC are among them. This illustrates how the composition of 3billion’s test aligns with an international research direction that is beginning to view childhood cancer predisposition genes through the lens of newborn screening. (👉See the genes covered by 3B-NEO)

This study is significant as the first to quantify, at a population level, that the horizon of genomic newborn screening can extend beyond metabolic and immunologic disorders to childhood cancer predisposition. At the same time, it makes clear that challenges remain—cost-effectiveness, implementation in clinical practice, and ethical considerations.

* This article is intended to introduce the principles and evidence behind such testing in general terms, not to provide individualized clinical interpretation for any particular patient. Whether and when testing is applied is determined by the judgment of the treating physician.

---

References

1. Diller, L., Cherkerzian, S., Cinelli, A. E., Housman, D., Gillani, R., Hamilton, K. V., Yeh, J. M., Bhattacharjee, A. & Parad, R. B. Population-based genomic detection of childhood cancer predisposition using newborn dried blood spots. Nature Communications, 2026. https://doi.org/10.1038/s41467-026-76296-8