The unique utility of newborn dried blood spots for driving discovery in childhood cancer research

Abstract

Background

Childhood cancer, though rare, remains the leading cause of disease-related death among children in the United States. The overall incidence has been rising for the past several decades, and the factors underlying the increasing rates are unclear. This unexplained increase underscores the critical need for high quality, well-powered epidemiologic research to elucidate causes and natural history of childhood cancers.

Content

Newborn screening (NBS) utilizes dried blood spots collected soon after birth to identify serious, treatable conditions and enable timely intervention, testing over 98% of infants born in the United States annually. Some NBS programs allow for the retention and release of residual dried blood spots (rDBS) for research. rDBS enable population-based assessment of both endogenous and exogenous factors including genetic, epigenetic, metabolic, immune characteristics, and environmental exposures during the perinatal period. This resource can overcome challenges of recall bias, exposure misclassification, and the impracticality of prospective cohort studies for rare cancers. Investigators have successfully used rDBS in epidemiologic studies, substantially advancing knowledge of several types of childhood cancers. Despite variability in storage and analyte stability, rDBS remain an invaluable resource to promote child health and constitute the only widely available pre-diagnostic biospecimen source for studies of childhood cancers.

Summary

rDBS from NBS programs represent a population-based, widely collected, pre-diagnostic biospecimen to investigate genetic, epigenetic, metabolic, and immune pathways underlying childhood cancer. Despite variability in storage policies, rDBS offer an unparalleled resource to advance etiologic understanding. Their integration into research could transform knowledge of childhood cancer causes and inform strategies for prevention and surveillance.

Introduction

Childhood cancer affects approximately 429,000 children annually worldwide, with survival rates varying dramatically between high-income countries (>80%) and lower-middle-income countries (<30%). (1) In the United States, an estimated 14,690 children and adolescents age <20 years will be diagnosed with cancer in 2025. (2) In contrast to adult cancers, which often originate from epithelial cells, childhood cancers most commonly originate from immature cell types including hematopoietic progenitors, embryonal cells, or central nervous system (CNS) progenitors. Likewise, the distribution of cancers observed in children is distinct; the most common diagnoses in children are leukemias, central nervous system tumors, and lymphomas. Given their unique characteristics, childhood cancers are classified by the International Classification of Childhood Cancer (ICCC-3), (3) which classifies tumors primarily by morphology rather than primary site.

Incidence

The incidence of childhood cancer and distribution of tumor types demonstrate marked variation across age (Figure 1). National data reveals relative differences in the incidence of childhood cancer overall, and major subtypes, across the United States (Figure 2). These differences may be due to variation in population demographic characteristics or environmental exposures and can inform etiologic hypotheses. Several investigations have also reported variation in age-specific incidence rates by demographic factors such as sex, race, and ethnicity. Across nearly all pediatric cancer types, male children exhibit higher incidence than females, a difference consistently observed across ages. (4) Racial and ethnic differences are also observed for most childhood cancers. (5) Although non-Hispanic White children have the highest overall childhood cancer incidence, Hispanic children exhibit higher incidence rates of acute lymphoblastic leukemia (ALL) and several solid tumors. Black children, by contrast, have a strikingly lower incidence of ALL and Ewing sarcoma, and a moderately lower incidence of several other subtypes. These patterns are not uniformly distributed across age but rather show age-specific incidence disparities that suggest underlying differences in risk factors such as genetic variation, environmental exposures, and social determinants of health.

Figure 1.

Distribution of Childhood Cancer Incidence by ICCC groupings and single year of age, NCCR 2017-2021

Figure 2.

Incidence Rate per Million by state (A) all childhood cancers combined; (B) leukemias; (C) non-CNS embryonal; (D) CNS; (E) bone and soft tissue sarcoma; (F) lymphomas

There has been a gradual but consistent increase in childhood cancer incidence over the past several decades. Between 2000-2019, the overall incidence in the United States has risen by approximately 0.8% annually. (6) This upward trend is most pronounced for certain tumor types, including non-Hodgkin lymphoma, hepatoblastoma, Ewing sarcoma, and thyroid carcinoma (Figure 3). Notably, increases have been observed across all major racial and ethnic groups, though the magnitude varies. There is also evidence of socioeconomic disparities in incidence trends. Children who reside in neighborhoods with lower socioeconomic status (SES) indicators have shown a disproportionately greater increase in incidence compared to those who live in higher SES neighborhoods. The factors underlying this increase are not fully understood, but have been hypothesized to include environmental exposures, changing prevalence of maternal and perinatal risk factors, and evolving population demographics. This unexplained increase underscores the critical need for high quality, well-powered epidemiologic research to elucidate causes and possible preventive measures for the overall rise in cancer incidence among children.

Figure 3.

Average Annual Percent Change (AAPC) Estimates (age < 20 years) by ICCC category, NCCR 2012-2021

Etiology

Childhood cancers are heterogeneous; although there are some risk factors that are shared among several tumor types, each tumor group has a distinct spectrum of established or suspected risk factors. (7) Indeed, there is evidence of etiologic heterogeneity even among molecular subtypes within tumor groups (8) and epidemiologic investigation has increasingly acknowledged the need for subtype-specific analyses. (7) Although the rarity of each cancer subtype renders this aspiration impractical in many settings, increasing availability of molecular subtyping (9) will enable continued advancement in molecular epidemiology.

Evidence indicates that several childhood cancer subtypes originate in utero, and molecular studies detecting initiating chromosomal rearrangements in blood collected at birth confirm a prenatal origin for several subtypes of acute leukemia. (10) This direct evidence, in combination with indirect evidence including the histologic and transcriptional resemblance (11) of embryonal tumor cells to immature fetal cells and the early onset of childhood cancers have naturally led to investigation of in utero and perinatal factors as potential causes of childhood cancers. Birth weight (12), structural birth defects (13), mode of birth (14), parental age (15), and maternal conditions during pregnancy (16) are among the perinatal factors investigated extensively.

Environmental exposures have been implicated as risk factors for several cancers but definitive conclusions are limited by challenges in exposure assessment. High-dose ionizing radiation is a confirmed cause of leukemia and some CNS tumors. (7) Prenatal and early-life exposure to pesticides, solvents, and traffic-related air pollution have been implicated as risk factors for certain tumor types, and some associations have strong evidence from meta-analyses, such as benzene exposure and AML.(17) Other potential environmental contributors include parental smoking , oil and gas exposure , heavy metals , per- and polyfluoroalkyl substances , though the literature on these exposures is not yet robust.

Chromosomal abnormalities, structural variation, and germline pathogenic variants markedly elevate risk but underlie a small proportion of childhood cancer cases. Well-established germline genetic risk factors for childhood cancer include Down syndrome, Li-Fraumeni syndrome, Beckwith–Wiedemann syndrome, DICER1 syndrome, Denys-Drash syndrome, RB1 mutation, WAGR syndrome, and neurofibromatosis type I, among others. (18) Common germline genetic risk variants have also been identified by genome-wide association studies (GWAS) of ALL, Ewing sarcoma, Langerhans cell histiocytosis, neuroblastoma, osteosarcoma, and Wilms tumor. (7)

Newborn Screening Practice and Culture

Newborn screening (NBS) in the United States is a public health program aimed at the early identification of serious but treatable conditions in newborns, enabling timely interventions that can prevent significant morbidity or mortality. The Wilson and Jungner criteria, established in 1968, have guided NBS program decisions for decades. (19) These principles posit that conditions screened should be early-onset, serious health problems with recognizable latent or early symptomatic stage and available treatment. Conditions should also have a screening test that is acceptable to the population and favorable cost benefit for early screening.

NBS is currently administered by 56 state and territorial programs, screening over 98% of the approximately 3.6 million U.S. newborns annually. The Recommended Uniform Screening Panel includes 38 core conditions and 26 secondary disorders for screening. The NBS process involves blood collection, typically within 24–48 hours after birth, via heelstick onto filter paper, transport to a state laboratory, analysis of the dried blood spot (DBS), and rapid follow-up for abnormal results. After screening is complete, residual dried blood spots (rDBS) are stored for confirmation of positive results or reanalysis, if needed. Most states also retain rDBS for use in development of new screening tests and some allow the release of rDBS for public health research, although retention time and storage practice vary widely by state (Table 1 and Figure 4), from 30 days to indefinite.

Table 1.

State rDBS retention time and storage conditions, among states with retention for at least one year

State/NBS Program	2023 births	Retention time (Years)	DBS Storage Conditions	Policy on Release of De- identified Specimens‡	Policy on Release for Research‡
California (91)	400,108	Indefinite	−20°C with desiccant	Yes, population-wide*	Release of de-identified specimens is allowed for public health research; release of identified specimens requires signed parental consent
Maine (92)	11,627	Indefinite	−20 °C	No	Written parental consent required
Michigan° (93)	99,124	35	−20 °C	Yes, subset	Release of de-identified specimens is allowed for public health research only if a parent opted in at the time of birth; release of identified specimens requires signed parental consent
Maryland (94)	65,594	25	4 °C	Yes, population-wide*	Release of de-identified specimens is allowed for public health research; release of identified specimens requires signed parental consent
Rhode Island (95)	9,805	23	NR	No	Not reported
Washington (96)	80,932	21	RT	Yes, population-wide*	Release of de-identified specimens is allowed for public health research; release of identified specimens requires signed parental consent
North Dakota (97)	9,647	18	RT	No	Written parental consent required
Massachusetts (98)	67,093	15	−20 °C	No	Written parental consent required
New York (99)	203,612	10	4 °C	Yes, population-wide*	Release of de-identified specimens is allowed for public health research; release of identified specimens requires signed parental consent
Iowa (100)	36,052	5	1 year at −80 °C then 4 years at RT	No	Written parental consent required
Minnesota (101)	61,715	5	−20°C with desiccant	No	Written parental consent required
Missouri (102)	67,123	5	−20 °C to −30 °C with desiccant	Yes, population-wide*	Release of de-identified specimens is allowed for public health research; release of identified specimens requires signed parental consent
North Carolina (103)	120,082	5	RT	No	Written parental consent required
Alaska (104)	9,015	3	RT	No	Written parental consent required
Connecticut (105)	34,559	3	−80 °C	No	Written parental consent required
New Jersey (106)	101,001	2	RT	No	Written parental consent required
Ohio (107)	126,896	2	−20 °F	No	Written parental consent required
Puerto Rico (108)	18,601	2	4 °C	Yes, subset	Release of de-identified specimens is allowed for public health research only if a parent opted in at the time of birth; release of identified specimens requires signed parental consent
Texas† (109)	387,945	2	RT	Yes, subset	Release of de-identified specimens is allowed for public health research only if a parent opted in at the time of birth; release of identified specimens requires signed parental consent
Idaho (110)	22,397	1½	RT	No	Written parental consent required
Oregon (111)	38,298	1½	RT	No	Written parental consent required
Arkansas (112)	35,264	1	−20°C with desiccant	No	Written parental consent required
District of Columbia (113)	7,896	1	RT	No	Written parental consent required
Hawaii (114)	14,808	1	RT	No	Written parental consent required
Mississippi (115)	34,459	1	NR	No	Not reported
Montana (116)	11,078	1	−20 °C	No	Written parental consent required
New Mexico (117)	20,951	1	RT	No	Written parental consent required
Pennsylvania (118)	126,951	1	−20 °C	No	Written parental consent required
South Carolina (119)	57,729	1	−20 °C	No	Written parental consent required
Tennessee (120)	83,021	1	2-8 °C	No	Written parental consent required
Vermont (121	5,065	1	−20 °C	No	Written parental consent required
Wisconsin (122)	59,754	1	4-8 °C	No	Written parental consent required

NR: not reported; RT: Room Temperature

^*States that allow release of de-identified specimens on a population-wide basis have provisions for parents to opt out by requesting destruction of their child's specimen; individuals who have reached the age of majority may also request destruction of their own rDBS

^°Michigan rDBS collected prior to May 1, 2010 may be released for research in a de-identified manner; all specimens collected on or after May 1, 2010 require a parent to opt-in for release of de-identified specimens

^†Texas rDBS collected on or after June 1, 2012 may be stored for up to 25 years if a parent opted in to storage and release of de-identified specimens for research use

^‡Data obtained from state legislative documents, state newborn screening program websites, and Association of Public Health Laboratories (123), accessed between September 29, 2025 and October 10, 2025

Figure 4.

Residual Dried Blood Spot Retention Times

Storage conditions, which can affect rDBS suitability for some downstream applications (20,21), range from ambient temperature, to temperature- and humidity- controlled environments such as −20°C freezers which preserve sample integrity for extended periods. Likewise, consent processes for DBS retention and release for research differ (Table 1), with some states adopting opt-in (explicit permission required) and others using opt-out (passive consent presumed unless declined) procedures. In most states, if consent for research release is not acquired as part of standard NBS procedures at birth, states will release rDBS only with written consent of the parents.

Uses of Residual Dried Blood Spots in Childhood Cancer Research

The potential of rDBS to advance childhood cancer research is immense. rDBS offer a unique and extremely valuable resource to assess genetic and epigenetic factors, as well as viral, nutritional, and environmental exposures present at birth, with possible extrapolation to earlier in utero exposures. There are numerous analytes detectable in rDBS reported to date, including indicators of endocrine, immune, reproductive, and metabolic function, nutritional status and infectious disease status, as well as environmental exposures such as benzene oxide, heavy metals, cotinine, pesticides, perfluorinated compounds, polychlorinated biphenyls, polybrominated diphenyl ethers, bisphenol A, phosphatidyl ethanol, and polyfluoroalkyl chemicals. In addition, availability of DNA and RNA enables genetic, epigenetic and transcriptomic research. Ongoing utilization of rDBS to elucidate genetic, environmental, and infectious exposures and biological response associated with pediatric cancer risk could enable primary prevention for modifiable factors. Furthermore, rDBS biobanks offer an invaluable biospecimen for direct measurement of early life exposures given that they are population-based and, when retention times are long, enable well-powered research for even very rare subtypes of cancer.

The resources that state-based biobanks of rDBS can provide are particularly valuable given the historical challenges of epidemiologic research for childhood cancer. The rarity of pediatric cancer subtypes, selection bias in case-control studies, and recall bias in questionnaire-based exposure assessments have limited scientific progress, particularly for very rare subtypes. Additionally, accurate assessment of many exposures is difficult to achieve without biospecimens (22) Self-reported exposure data collected via questionnaire gives rise to the potential for both differential (e.g., due to recall bias) and non-differential (e.g., due to inaccurate recollection of exposures that may have occurred many years prior to data collection) exposure misclassification. By contrast, biospecimens can provide accurate, unbiased estimates of exposures through the use of biomarkers. However, they must be collected prior to the onset of disease for use in etiologic research; when collected contemporaneously with diagnosis, the disease process and subsequent anticancer treatment may alter exposure, biomarker status, or both (e.g., reverse causality). Due to the rarity of pediatric cancers, well-powered cohort studies for prospective collection of data and biologic specimens are not economically or logistically feasible, particularly for the study of rare tumors (e.g., hepatoblastoma) or molecularly defined subtypes of more common tumors (e.g., cytogenetic subtypes of ALL). Therefore, rDBS constitute the only population-based source of pre-diagnostic biospecimens. In addition, state biobanks store rDBS for the entire birth cohort of each year within their retention time, and thus can provide population-based samples from representative controls. Birth certificates can provide data on basic demographic characteristics as well as some exposures of interest (e.g., parental age, birth weight, gestational age, pregnancy complications, mode of birth, and birth defects, among others), while state cancer registries provide diagnostic data for cases.

For researchers to effectively leverage rDBS for the study of childhood pediatric cancers, there are several logistical and administrative barriers that must be overcome. In the United States, only thirteen states routinely store rDBS for more than five years (Figure 4 and Table 1). At the time of this publication, five states (California, Maryland, Missouri, New York, and Washington) store all rDBS and allow release of de-identified specimens for public health research purposes. Another two states and one territory (Michigan, Texas, and Puerto Rico) obtain consent for rDBS release for research purposes at the time of birth and will release de-identified specimens for this subset. The remaining states require signed parental consent in order to release rDBS for public health research.

De-identified rDBS-based linkage studies of childhood cancer are possible through processes defined by each state and require ethics and regulatory approvals, as well as adherence to strict guidelines for data and specimen security. For approved studies, the process generally follows as such: cases of interest are identified by the state’s cancer registry and linked to the state’s birth registry to identify which cases were born in that state; for case-control studies, controls are then randomly selected based on pre-determined criteria; information on eligible cases and controls is then sent to the biobank for specimen retrieval. De-identified datasets and specimens are then prepared and sent to investigators.

Investigators who wish to obtain specimens from states that do not allow for de-identified release complete a similar process of obtaining regulatory, scientific, and ethical approvals. Case identification and contact procedures may vary and may involve rapid ascertainment through hospitals or organizations such as the Children’s Oncology Group, or through the state’s cancer registry.

There are constraints for rDBS-based research, either due to state policies or other inherent limitations. Perhaps most critically, children who are diagnosed after the retention time for their rDBS has ended will never have the opportunity to be represented in this type of research. This is particularly important when considering that the distribution of cancer varies by age at diagnosis (Figure 1); investigation of cancers that occur most commonly among older children and adolescents is limited to the few states with retention of more than 10 years. Additionally, the data available through de-identified linkage studies, which generally do not involve participant contact, is limited to that available from birth and cancer registries. Many state programs also have restrictions on access to certain data elements or on the granularity of de-identified birth certificate data which may be released. Tumor specimens and detailed somatic molecular classification, which are increasingly important for classifying tumors and characterizing etiologic heterogeneity, are also unavailable. Data availability challenges can be overcome with contact and active recruitment of childhood cancer cases and their families into research studies. In this setting, participants can provide questionnaire data and consent for researchers to obtain stored tumor specimens from their treatment institution. However, these efforts require substantial time and financial resources, and non-response bias may be a concern if participation rates are low.

Despite these challenges, rDBS remain a valuable resource and a growing body of research has utilized rDBS to elucidate the etiology of childhood cancers. Here we summarize areas of investigation that are most robust within the childhood cancer literature. These applications are described below and select prior work utilizing rDBS for childhood cancer etiologic research is summarized in Table 2.

Table 2.

Select publications utilizing rDBS for childhood cancer research

First author	PMID	Year	Method	Exposure	Outcome	rDBS specimen size
JS Chang(73)	21653647	2011	Luminex bead-based assay	IL2, IL4, IL5, IL6, IL10, IL12, IL13, IL17, GM-CSF, IFN-γ, TNF-α	ALL	⅛ portion of each DBS
AP Chokkalingam(124)	23576692	2013	Lactobacillus casei microbiologic growth assay	HbFol	ALL, AML	NR
AM Dahlin(125)	26290144	2015	Illumina HumanOmni2.5-8 BeadChip	Single nucleotide variants in genes CCND2, CTNNB1, DDX3X, GLI2, SMARCA4, MYC, MYCN, PTCH1, TP53 and MLL2	Medulloblastoma	Two x 3.2mm
P Bhatti(126)	25348494	2016	LC-MS/MS	25(OH)D2, 25(OH)D3	CNS tumors	6.35mm
G Bogdanovic(127)	27552439	2016	Illumina MiSeq Sequencing System	Viromics	ALL	Four x 3mm
SS Francis(72)	27979823	2017	ddPCR	CMV and EBV	ALL	¼ portion of each DBS
LM Morimoto(128)	29475970	2018	LC-MS/MS	Androstenedione, DHEA, Estradiol, Estriol, Estrone, Testosterone, Progesterone	Testicular germ cell tumor	Four x 3mm
A de Smith(129)	29923177	2018	Affymetrix Axiom World LAT Array	BMI1 polymorphism	ALL	Three x 3.0mm
JL Wiemels(29)	29348612	2018	Affymetrix Axiom World LAT Array	Single nucleotide variants	ALL	Three x 3.0mm
SH Soegaard(74)	30217873	2018	multiplex sandwich immunoassay (Meso-Scale)	IL6 and its soluble receptor sIL6Ra, IL8, IL10, IL12, IL17, IL18, TGFb1, monocyte chemotactic protein (MCP)-1, and C-reactive protein (CRP)	ALL	Two x 3mm
AL Brown(130)	31350265	2019	Illumina Infinium Global Screening Array	Single nucleotide variants	ALL among children with Down syndrome	Six x 3mm
LM Petrick(54)	30904619	2019	Untargeted LC-HRMS metabolomics	Untargeted metabolomics	ALL	4.7mm
AB Nielsen(75)	30923090	2019	Enzyme-linked immunosorbent assay	Arginase 2	ALL	⅓ portion of each DBS
C Zhang(31)	33115534	2020	Affymetrix Axiom World LAT Array	Polygenic risk score	Ependymoma	⅓ portion of each DBS
Y Yano(64)	31760269	2020	nLC-HRMS	Untargeted adductomics	ALL, AML	4.7mm
TP Whitehead(82)	34078642	2021	R&D Systems Multiplex Cytokine Assay Kit	IL1, IL4, IL6, IL8, IL10, IL12p70, GM-CSF, TNFα, VEGF	ALL	Two x 4.7mm
LM Petrick(56)	33971561	2021	Untargeted LC-HRMS metabolomics	Untargeted metabolomics	AML	4.7mm
EM Nickels(41)	36178055	2022	Illumina Infinium Methylation EPIC BeadChip	DNA methylation	ALL	¼ portion of each DBS
S Li(131)	36070312	2022	Precision Medicine Diversity Array	Local ancestry	Pilocytic astrocytoma	⅓ portion of each DBS
MA Richard(47)	36226634	2022	Illumina Infinium Methylation EPIC BeadChip	DNA methylation	Lymphoma
K Xu(39)	35717575	2022	450K DNA methylation arrays or EPIC arrays; Illumina Human OmniExpress V1 platform and the Affymetrix Axiom World LAT Array	DNA methylation and germline genetic variants	ALL	¼ section of each DBS for genotyping, and ¼ section of each DBS for methylation arrays
D He(132)	36730579	2023	Untargeted LC-HRMS metabolomics	Cotinine and hydroxycotinine	Retinoblastoma	NR
AE Janitz(60)	38104918	2023	Gas chromatography-tandem mass spectrometry	trans-nonachlor, p,p’-dichlorodiphenyldichloroethylene (p,p’-DDE), p,p’-dichlorodiphenyltrichloroethane (p,p’-DDT), (β) hexachlorcyclohexane (β-HCCH), and hexachlorobenzene (HCB); PCB congeners 118, 138, 153 and 180; PBDE congeners 47, 99, 100, 85, 153, 154; and PBB-153	AML	Whole spot
C Metayer(55)	36831356	2023	LC-HRMS with HILIC chromatography	Folate pathway metabolites	ALL	4.7mm
Q Yan(57)	38130370	2023	Untargeted LC-HRMS metabolomics	Untargeted metabolomics	Retinoblastoma	5mm
Y Chen(59)	37866539	2024	Untargeted LC-HRMS metabolomics	PFAS	Retinoblastoma	5mm
DM Gianferante(133)	37596165	2024	Illumina Infinium Global Screening Array	Polygenic risk scores	Osteosarcoma	Three x 3.0mm
A Ghantous(40)	39443995	2024	Illumina HM450 array or EpiTyper	DNA methylation	ALL	¼ portion of each DBS or Six x 3mm
LM Morimoto(58)	40716601	2025	untargeted LC-HRMS	PFAS	ALL	4.7mm
T Xie(33)	40465396	2025	Illumina Infinium Global Screening Array	Predicted gene expression	Hepatoblastoma	Two x 3mm

NR: Not reported; ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; ddPCR: droplet digital polymerase chain reaction; GM-CSF: granulocyte-macrophage colony-stimulating factor; TNFα: tumor necrosis factor alpha; VEGF: vascular endothelial growth factor; LC-MS/MS: liquid chromatography - tandem mass spectrometry; nLC-HRMS: nano-liquid chromatography high resolution mass spectrometry; HILIC: hydrophilic interaction; CMV: cytomegalovirus; EBV: Epstein-Barr virus

Germline Genetics

rDBS are frequently used to derive DNA sufficient to conduct array genotyping or standard-read whole exome or whole genome sequencing. Quality and quantity of DNA from rDBS depends on storage condition (23) as well as method of extraction. (24) However, DNA yield and genotyping array or sequencing success do not appear to correlate with age of rDBS or filter paper type. (25) This is perhaps not surprising given that DNA’s stability may last millennia. (26) Typical yields for a 6mm punch range from tens to hundreds of nanograms with acceptable purity. (27) To our knowledge no one has attempted long-read whole genome sequencing in rDBS, despite this technology’s growing utility, likely due to its requirement for high-mass (>3 micrograms) and unfragmented DNA, which is difficult if not impossible to meet with rDBS. (28)

DNA derived from rDBS have been successfully used to conduct very large-scale genotyping assays underlying GWAS for pediatric ALL (29) and glioma (30), with smaller-scale studies of ependymoma (31) and osteosarcoma (32) also found in the literature. Genome wide single nucleotide polymorphism (SNP) array data from rDBS has also been used as part of a transcriptome-wide association study (TWAS) of hepatoblastoma. (33) While we have not identified large-scale sequencing studies of childhood cancer based on rDBS-derived DNA, targeted sequencing of cancer susceptibility genes using rDBS was recently reported. (34) It is also worth noting that newborn whole genome sequencing is becoming a possibility with the recent publication of several pilot studies. (35) Newborn screening for pathogenic and likely pathogenic variants in cancer susceptibility is being explored in epidemiologic and simulation studies. (34,36) There is already screening for a specific TP53 variant (p.R337H) that is common in Parana state in Brazil, which predisposes to adrenocortical carcinoma. (37)

Methylation

The most frequently examined epigenetic alteration in cancer epidemiology studies is DNA methylation, (38) largely due to the availability of relatively low-cost and high-throughput DNA methylation arrays. rDBS can be used as a source of DNA for assessing variation in DNA methylation, and while comparatively few studies have leveraged rDBS for epigenetics research in childhood solid tumors, there have been several studies focused on childhood ALL.

In the largest epigenome wide association study (EWAS) of ALL, variation in DNA methylation at a CpG in the ARID5B gene was explained entirely by the known ALL GWAS SNPs in the region, while neonatal hypermethylation at the IKZF1 promoter region was found to partially mediate the effects of a nearby ALL risk SNP. (39) A more recent EWAS identified significant hypermethylation at the imprinted gene VTRNA2-1 as a potentially novel risk factor for childhood ALL. (40) In another case-control study of 41 monozygotic twins discordant for ALL, analysis of DNA methylation array data from rDBS revealed several loci and global hypomethylation associated with ALL risk. (41)

Other epigenetic studies of childhood ALL have leveraged the ability to use methylation profiles at birth to examine potential biomarkers of exposures, most notably for prenatal tobacco smoke exposure. (42) Case-control studies have not observed statistically significant associations between rDBS-derived epigenetic biomarkers of maternal smoking during pregnancy and childhood ALL risk, (39,43) supporting results from prior epidemiologic data. However, case-only analyses of rDBS methylation have revealed the potential role of tobacco exposure in the generation of somatic driver gene deletions in ALL patients, suggesting possible subtype-specific effects. (44)

Finally, rDBS have been leveraged to study the impact of congenital anomalies on epigenetic profiles at birth. rDBS-derived DNA methylation data from newborns with and without Down syndrome (DS), which is associated with an up to 30-fold increased risk of childhood ALL, (13) have revealed differentially methylated regions across the genome, the most significant of which overlapped RUNX1, a gene that is frequently altered in childhood ALL. (45) rDBS-derived DNA methylation data was also used to infer blood cell proportions in newborns with DS, and found a significantly increased proportion of B cells in DS-ALL cases than in DS controls. (46) Finally, a study of newborn DNA methylation in rDBS from children with both congenital heart defects and lymphoma identified potential novel risk factors for these associated conditions. (47)

More research on epigenetic variation in rDBS and childhood cancer risk is warranted, in particular in malignancies beyond ALL, to gain knowledge on potentially modifiable risk factors and on strategies for early detection. Identification of robust epigenetic biomarkers of intrauterine exposures beyond maternal smoking would present further opportunities for studying the causes of childhood cancer.

Metabolomics

Metabolomics refers to the measurement of endogenous metabolites, hormones, and signaling molecules in biological samples. rDBS have successfully been used as source material for both targeted and untargeted metabolomics (48,49) and these studies have described both endogenous (e.g., inborn errors of metabolism) and exogenous (e.g., environmental contaminants) sources of metabolic perturbation. Methods and workflows specific to rDBS have been successfully developed and validated; (50,51) generally, metabolic profiles derived from rDBS have been shown to be highly concordant with those derived from more traditional plasma samples. (50,52,53)

A growing number of pediatric cancer studies have leveraged rDBS for metabolomics. Multiple studies have investigated the potential role of nutrition-related metabolic perturbations in pediatric leukemia. (54,55) Another study of rDBS from acute myeloid leukemia cases and healthy controls identified multiple sex-specific metabolic predictors of disease. (56) An untargeted case-control study of retinoblastoma identified that disease risk was associated with perturbations to inflammatory and energy metabolism pathways at birth. (57) Other case-control studies have identified that the concentrations of perfluoroalkyl substances (PFAS) measured in rDBS were associated with increased ALL and retinoblastoma risk. (58,59)

The uses of rDBS in metabolomics for pediatric cancer have favored untargeted and discovery-based methods to date and few studies have applied targeted methods to identify how the presence of exogenous metabolites can influence pre-cancer indicators or overall risk. Janitz et al. conducted a nested case-control study of AML using rDBS to evaluate multiple persistent organic pollutants through targeted metabolomics, including organochlorine pesticides, polychlorinated biphenyls (PCB), polybrominated biphenyl (PBB), and polybrominated diphenyl ethers (PBDE), with mixed results. (60)

Metabolite measurement is a particularly promising tool for identifying the potentially putative chemical exposures in cases where complex mixtures are involved (e.g., perfluoroalkyl compounds). Such studies could also evaluate the relationships between environmental stressors and the presence of pre-cancer metabolic indicators for cancers with prenatal origins, such as leukemia. Continued advancement of technology and deeper understanding of methods to optimize high-quality, reproducible data from rDBS will undoubtedly further enhance the utility of this powerful tool.

Protein Adductomics

Protein adducts, chemical modifications formed when reactive compounds bind to amino acid residues in proteins, provide another biomarker useful for assessment of in utero exposures. Protein adductomic assays systematically detect and quantify covalent modifications on proteins that result from exposure to endogenous metabolites, such as reactive oxygen species (61) or exogenous chemicals, such as environmental pollutants (62) or other bioactive compounds such as anticancer drugs. (63)

Prior literature on protein adductomics in childhood cancer is limited. One publication reports that T-cell ALL cases had a higher abundance of reactive carbonyl species compared to controls and that rDBS of acute myeloid leukemia cases had lower abundance of a Cys34 disulfide of homocysteine. (64) Despite the limited prior research utilizing rDBS to study the associations between the protein adducts of environmental and other endogenous or exogenous exposures, there is broad recognition of the potential of this field to substantially contribute to our knowledge of how in utero exposures impact childhood cancer risk. (49,65)

Immune Profiling

Biomarkers of immune dysregulation can be used to assess neonatal health and risk of disease later in childhood, and the use of rDBS to measure immune biomarkers at birth is well established. Studies typically incorporate multiplexed assays to measure various cytokines and immune-related proteins, (66) maximizing the relatively small amount of starting material in rDBS. In this way, rDBS have been used to identify biomarkers of congenital heart disease (67) and to investigate the association between neonatal immune dysregulation and subsequent risk of developmental delay (68) and schizophrenia in later life. (69) rDBS have also been used to study the effects of maternal and infant genetics on neonatal immune mediators. (70)

In childhood cancer, the use of rDBS to measure immune-related proteins have been limited to studies of ALL. Immune dysregulation at birth is of particular interest in childhood ALL, given the proposed role of early-life infections in ALL etiology (71) and the discovery of perinatal risk factors that may modulate immune development such as planned Cesarean section (14) and in utero cytomegalovirus (CMV) infection. (72) In a first study of its kind, lower neonatal levels of the anti-inflammatory cytokine IL-10 measured in rDBS were found to be significantly associated with later risk of ALL. (73) Although subsequent studies have struggled to measure this cytokine above the limits of detection, rDBS-measured variation in other inflammatory markers have been associated with childhood ALL risk. (74,75) In another case-control study, increased measures of the CMV protein CMV-IL-10, which is known to interact with the human IL-10 receptor and was inversely correlated with human IL-10 levels, were associated with increased risk of childhood ALL. (76) Further studies of neonatal immune markers and risk of other childhood cancers are warranted, given the potential role of aberrant immune surveillance in the development of hematological and solid tumors.

Considerations for rDBS Use in Downstream Applications

There are important considerations for the use of rDBS in genomic, epigenomic, and exposomic applications. Although it is out of scope for this manuscript to provide a comprehensive review of the impact of storage duration and conditions on feasibility for individual downstream assays and a discussion of each assay’s requirements for blood volume or input mass, thorough reviews on these topics have been previously published. (21,27,49) Here we highlight some key considerations.

DNA extracted from rDBS are consistently of sufficient quality and quantity for a wide range of downstream applications. Therefore, it is feasible to generate genetic and epigenetic data from rDBS using most DNA-based array and sequencing platforms regardless of storage duration or conditions. However, we note that technologies that require high input mass or unfragmented DNA, such as long-read sequencing, described above, are far less likely to be successful when deployed in rDBS.

Although RNA-based assays have not yet been widely utilized in rDBS-based research on childhood cancer, it is worth noting evidence that sufficient RNA can be extracted from rDBS to conduct quantitative PCR (qPCR), array gene expression profiling, and RNA sequencing. (77,78) Furthermore, there is evidence that, among samples stored between 1 to 20 years at 4°C, amplifiable transcripts are present even in the oldest samples and that RNA degradation was not substantial over time. (79) In a comparison of rDBS stored for 8 to 10 years in either frozen or ambient temperature conditions, array gene expression data showed strong correlation between storage conditions, although fewer genes were expressed in unfrozen specimens. (80) While these results are promising, sample sizes in each study were small; comprehensive characterization of rDBS suitability for RNA-based applications are necessary to inform future work.

In considering rDBS as the source material for measurement of metabolites or protein adducts, analyte stability, which may vary by storage conditions, is a primary concern. (48,49,81) However, an evaluation of metabolite stability in rDBS from 899 healthy children, collected between 1983 and 2011, reported that endogenous metabolites measured remained stable over the years tested and there were no significant trends associated with storage duration. (48) Overall, if storage methods are consistently temperature- and humidity-controlled, the evidence suggests that many metabolites will remain stable in rDBS even over decades of storage. Therefore, metabolomic work using rDBS is feasible but may be limited to rDBS from states with optimal storage conditions for this downstream application.

In research focused on immune profiling, there is wide variation in the half-life of different immune biomarkers, such that some proteins are largely undetectable or at very low levels in rDBS, and different approaches should be considered to account for measures below the limit of detection. (73,82) Furthermore, the stability and hence detectability of immune markers may vary based on the storage time of rDBS, the number of freeze-thaw cycles, and the temperature at which bloodspots are stored long-term. (83)

Future Directions

Future work utilizing assays previously deployed to study other pediatric conditions, such as transcriptomics and protein adductomics, will continue to advance the field. There are also clear directions for the future of childhood cancer research which are unique to this field due to the in utero origin of disease and, among leukemias, demonstration that pre-malignant cells are detectable in birth samples.

Translocation screening

Backtracking studies, which trace patient-specific translocations from ALL diagnosis back to rDBS, (10) confirm the prenatal origin of acute leukemia and establish that pre-leukemia, defined as the presence of leukemia-driving translocations in the absence of overt disease, can be utilized as a screening and risk prediction tool for childhood leukemia.

There are several obstacles to realizing this possibility. Classically, detection of leukemic translocations in rDBS requires knowledge of the exact translocation breakpoint, which is unique to each individual. Translocation screening methods scalable population-wide would need to be agnostic to the exact DNA breakpoint. GIPFEL (Genomic Inverse PCR for Exploration of Ligated Breakpoints) is one such method, although it requires B-cells isolated from cord blood. (84) As technology advances, the ability to detect pre-leukemia using methods adapted specifically for rDBS is inevitable. This transformative milestone will enable study of the natural history of pre-leukemia and prospective follow up of infants with pre-leukemia to identify post-natal factors associated with disease development.

Twin tracking

The in utero origin of childhood leukemia was first posited after examining twins with concordant leukemia that featured identical translocation breakpoints, which is unlikely to have occurred by chance (85). Instead, it appeared that the translocations occurred in one twin and spread to the other through shared placental vasculature, which occurs in ~60% of monozygotic twins. Although risk of leukemia in unaffected monozygotic twins may be lower than previously thought, it is possible to track pre-leukemia in these individuals based on their affected twin’s specific translocation breakpoint. (41) rDBS could be used to establish a baseline for “twin tracking” and, for complementary research questions, to select twins for prospective sample collection, based on pre-leukemia concordance status at birth.

cfDNA

In addition to the proposed use of rDBS for newborn screening of preleukemic translocations, the possibility for early detection of pediatric solid tumors using neonatal blood samples remains unexplored. There is evidence that circulating tumor DNA (ctDNA) can be detected in diagnostic plasma samples from children with solid tumors using cell-free DNA (cfDNA), (86) and that plasma cfDNA can even be used for early cancer detection in high-risk children such as those with Li-Fraumeni syndrome. (87) Whether pre-diagnostic ctDNA may be detectable in cfDNA in rDBS should be investigated for childhood solid tumors, particularly in newborns harboring pathogenic germline variants in cancer predisposition genes and for cancers that typically present early in life (e.g., neuroblastoma), as earlier detection may improve patient outcomes.

Conclusion

Childhood cancer is the leading cause of death from disease among children in the United States. Additionally, incidence is rising, and survivors often suffer lifelong late effects due to the toxic treatment received. Due to its widespread collection prior to disease onset, rDBS offer an ideal biospecimen and rDBS-based research has the potential to revolutionize our understanding of the causes and natural history of childhood cancers. The American College of Medical Genetics and Genomics (ACMG) describes rDBS as a “valuable national resource that significantly contributes to the health of all children and families”. (88) Substantial progress in elucidating the genetic, molecular, and environmental epidemiology of childhood cancer in the last several decades, and knowledge gained in whole or in part from rDBS-derived data has contributed significantly to these advances. Continued integration of rDBS into epidemiologic research will enable high-impact investigation of the causes of childhood cancer and ultimately facilitate better outcomes through strategies for early detection, risk prediction, surveillance and prevention. (89,90)

State policies on rDBS retention time, storage conditions, and release for public health research have substantial impact on researchers’ ability to realize the full potential of rDBS-based research. In addition, under current circumstances, prior and future work will inherently lack nationally representative samples due to the short retention time of many state’s specimens, preventing their inclusion in research. To improve generalizability of research results and ensure that children nationwide may benefit from insights gained, it is imperative for stakeholder communities to advocate for state policies that recognize the unparalleled scientific and public health value rDBS hold while balancing informed public participation, transparency, confidentiality and data security. Indeed, the ACMG states that “As states continually revise policies affecting retention of residual dried blood spots, careful consideration must be given to the irreplaceable value to child health provided by these specimens.” (88)

Nonstandard Abbreviations

NBS

Newborn Screening

rDBS

Residual Dried Blood Spots

CNS

Central Nervous System

ICCC-3

International Classification of Childhood Cancer, Third Edition

SES

Socioeconomic Status

ALL

Acute Lymphoblastic Leukemia

GWAS

Genome-Wide Association Studies

DS

Down Syndrome

ACMG

American College of Medical Genetics and Genomics

AML

Acute Myeloid Leukemia

SNP

Single Nucleotide Polymorphism

EWAS

Epigenome-Wide Association Study

TWAS

Transcriptome-Wide Association Study

PFAS

Per- and Polyfluoroalkyl Substances

PCB

Polychlorinated Biphenyl

PBB

Polybrominated Biphenyl

PBDE

Polybrominated Diphenyl Ether

CMV

Cytomegalovirus

IL-10

Interleukin-10

CMV-IL-10

Cytomegalovirus-encoded Interleukin-10 homolog

qPCR

Quantitative Polymerase Chain Reaction

cfDNA

Cell-Free DNA

ctDNA

Circulating Tumor DNA

GIPFEL

Genomic Inverse PCR for Exploration of Ligated Breakpoints

Human Genes:

Gene Symbol

HGNC Approved Name

TP53

Tumor Protein P53

ARID5B

AT-Rich Interaction Domain 5B

IKZF1

IKAROS Family Zinc Finger 1

VTRNA2-1

Vault RNA 2-1

RUNX1

Runt-Related Transcription Factor 1

IL10

Interleukin 10

RB1

RB Transcriptional Corepressor 1

DICER1

Dicer 1, Ribonuclease III