Mutational Signatures in Pediatric Hepatoblastoma: Molecular Mechanisms and Prognostic Potential

Gustavo Dib Dangoni¹, Talita Ferreira Marques Aguiar^2*, Guilherme Henrique Souza Bomfim³, Ana Cristina Victorino Krepischi^1*

¹Department of Genetics and Evolutionary Biology, Institute of Biosciences, Human Genome and Stem-Cell Research Center, University of São Paulo, São Paulo, Brazil

²Department of Pathology, Texas Children's Hospital, Texas, USA

³Department of Molecular Pathobiology, New York University College of Dentistry. New York, USA

---

Abstract

Cancer genomes reflect both the accumulation of somatic mutations and the underlying mutational processes that generate them. While somatic mutation burden provides a quantitative measure of genomic alterations, mutational signatures capture the mechanistic footprints of endogenous and exogenous DNA damage and repair pathways. This distinction is crucial in pediatric cancers, which typically exhibit low mutation burden arising from biologically distinct, developmentally constrained mechanisms. Recent large-scale genomic studies demonstrate that pediatric tumors harbor a restricted repertoire of COSMIC mutational signatures, in contrast to findings in adult cancers, which exhibit broader mutational diversity driven by cumulative environmental exposures. In this review, we systematically revisit, synthesize, and critically re-evaluate the current evidence on mutational signatures in pediatric malignancies, with a focus on hepatoblastomas.

---

Background

Understanding cancer evolution requires distinguishing between two overlapping but distinct molecular measures: somatic mutation burden, the total number of mutations accumulated in a tumor, and mutational signatures, which reflect the underlying mutational processes that generate these mutations. While mutational burden provides a quantitative assessment of genomic instability¹, mutational signatures offer mechanistic insights by revealing the characteristic patterns of DNA changes driven by specific causative agents such as environmental exposures, DNA repair deficiencies, or endogenous cellular processes^2-4. Mutational signatures are systematically catalogued in the COSMIC (Catalogue of Somatic Mutations in Cancer) repository (https://cancer.sanger.ac.uk/signatures/ - v3.5), which provides reference definitions for single-base substitution signatures (SBS), double-base substitution signatures (DBS), insertion/deletion signatures (ID), copy number signatures (CN), structural variations signatures (SV), and RNA single base substitution signatures (RNA SBS).

In pediatric cancers, this distinction is particularly important: children's tumors harbor fewer mutations than those in adults, and the composition and patterns of these mutations reveal distinct biological processes that drive early-onset tumorigenesis. Despite this difference, pediatric cancers are predominantly characterized by oncogenic gene fusions and other structural variants compared with adult cancers⁵, as well as by disruptions in the epigenetic machinery, such as alterations in DNA methylation and histone modifications⁶. This review examines both the overall mutational landscape and the specific mutational signatures that characterize pediatric malignancies, with particular focus on hepatoblastoma as a case study of how signature analysis can inform prognostic stratification and therapeutic approaches.

The molecular architecture of cancer generally reflects the age of onset¹. While adult malignancies result from accumulated genetic and environmental insults over decades, pediatric cancers emerge in a fundamentally different biological context, characterized by developmental processes and limited lifetime environmental exposures. This distinction is reflected in the mutational profiles of these tumors: notably, somatic mutation burden is typically much lower in pediatric cancers compared with adult malignancies. In children, structural variants, including copy-number alterations, and epigenetic modifications predominate over single-nucleotide variants⁷, whereas adult cancers accumulate mutations through both endogenous aging processes and exogenous exposures over longer biological spans. Figure 1 provides an overview of the developmental timing of these mutational processes, contrasting prolonged accumulation of environmental and age-related signatures in adults with the compressed, developmentally constrained mutational landscape characteristic of pediatric cancers.

Large-Scale Studies in pediatric mutational signatures

In this review, we evaluated the PeCan platform⁸, two main pan-cancer studies on mutational signatures in pediatric and young adult cancers (ages ranging from 0-25 years)^7,9, and all articles on hepatoblastomas that studied this feature. The most recent large-scale study by Thatikonda and colleagues (2023)⁹ examined 785 whole-genome tumors across 27 molecularly defined subtypes, providing the most comprehensive assessment of mutational signature diversity in childhood malignancies to date. This work was the second pan-cancer study to analyze mutational signatures in pediatric patients (younger than 18 years old) and young adults (up to 25 years old), revealing that pediatric tumors harbor a markedly restricted repertoire of mutational signatures relative to adult cancers, a finding with two particularly significant implications. First, homologous recombination repair–deficient (HRD) signatures, which are prevalent in adult breast and ovarian cancer, were notably underrepresented in pediatric cohorts, highlighting a fundamental biological distinction in DNA repair capacity and in the mutational processes that underpin tumorigenesis in children compared to adults. Second, the study identified a novel indel signature in pediatric leukemias, characterized by long insertions in non-repetitive regions, suggesting disease-specific mutational processes not previously recognized. These events were primarily detected in intronic and intergenic regions, as well as in exons of known cancer-related genes. These findings collectively establish a comprehensive reference map of COSMIC v3 mutational signatures in pediatric cancers, providing new insights into their biology with direct implications for molecular classification, biomarker discovery, and therapeutic stratification.

A key framework for interpreting pediatric mutational signatures derives from studies of clock-like mutational processes in normal somatic cells, driven by relatively constant mutational rates that correlate with chronological age. One such process is the spontaneous or enzymatic deamination of 5-methylcytosine to thymine, which is proposed as the etiology of the SBS1 signature¹⁰. Further analyses across 29 cell types demonstrated that SBS1 and SBS5/40 are ubiquitous, although their accumulation rates vary by tissue, reflecting steady age-related mutation accrual¹¹. Subsequent mechanistic modeling clarified that these signatures arise from regular cellular turnover or consistent DNA damage and repair cycles, a distinction that can be resolved by contrasting fast- versus slow-cycling cell lineages¹².

Notably, pediatric cancers exhibit these clock-like signatures (SBS1, SBS5, and indel signatures ID1/ID2) despite their dramatically shortened timeframe for mutation accumulation, although the overall number of operative mutational processes is lower than in adult tumors⁹. Beyond these endogenous clock-like patterns, additional recurrent features in pediatric cancers include the frequent presence of ROS-associated SBS18, late-replication errors (SBS8) rather than homologous recombination deficiency, and more restricted APOBEC activity (SBS2/SBS13) compared to adults.

SBS18, a reactive oxygen species (ROS)–associated mutational signature commonly observed in neuroblastoma¹³, has also been consistently reported across multiple hepatoblastoma cohorts, including independent studies^14-16 and analyses from the PeCan database⁸. Consistent with observations in liver and other high-metabolism tissues¹⁷, the recurrence of SBS18 across several pediatric cancers, including hepatoblastoma, likely reflects the intrinsically high oxidative metabolism of these tissues, in which elevated ROS levels mediate characteristic DNA damage. While not exclusive to hepatoblastoma, the enrichment of SBS18 reinforces the role of metabolic turnover as a key determinant of mutational processes in pediatric tumors arising from energetically demanding tissues.

Exogenous mutational imprints, although generally uncommon in pediatric cancers, can appear in context-dependent settings, as exemplified by the unexpected detection of the UV-light signature SBS7 in hypodiploid B-cell acute lymphoblastic leukemia, highlighting potential environmental or lineage-specific vulnerabilities⁹. Finally, the prevalence of HRD signatures remains lower than previously suggested, reinforcing that HRD-driven mutagenesis is not a major feature of childhood cancer biology⁹.

Adult Vs. Pediatric Signature Diversity

The contrast with adult cancers is striking in both scope and complexity. Aging can increase the total number of mutations in adults in a cumulative fashion, even with minimal environmental exposure, due to endogenous processes^1,2,4,10,18; for this reason, aging is a major risk factor for cancer. Nevertheless, the complexity of oncogenesis is mainly driven by exogenous exposures throughout life, such as tobacco smoking, UV radiation, and other carcinogens¹⁸, which accumulate DNA mutations.

The PCAWG analysis of adult tumors identified 49 single-base substitution signatures, 11 DBSs, and 17 indel signatures, a diversity exceeding that in pediatric malignancies. This expanded signature repertoire reflects the cumulative impact of lifelong exposure to a broad range of both endogenous processes and environmental factors⁴. Environmental factors significantly influence the mutational landscape of many adult cancers. Tobacco smoke (SBS4, DBS2) is the predominant driver of lung and head and neck carcinomas, while ultraviolet radiation (SBS7, DBS1) drives skin cancers. Additionally, APOBEC-mediated mutagenesis (SBS2/13) contributes broadly across multiple carcinoma types. Beyond environmental exposures, adult tumors frequently show a higher prevalence of HRD-associated SBS3 in breast and ovarian cancers, alongside abundant therapy-associated signatures, including those linked to platinum-based chemotherapy (SBS31, SBS35) and thiopurines (SBS32). The COSMIC database comprehensively cataloged these patterns⁴. Emerging evidence beyond classical exposures suggests that previously unconsidered exposures may also contribute to mutagenesis, as exemplified by the recent classification of ultra-processed foods as probable carcinogenic agents¹⁹, raising questions about dietary components, additives, contaminants, and processing-derived compounds generate distinct mutational signatures in malignancies, a hypothesis awaiting empirical validation.

A comparison can be drawn by combining these findings. Pediatric tumors are characterized by markedly lower overall mutation loads, fewer active mutational signatures, dominance of endogenous clock-like processes alongside reactive oxygen species–linked damage, restricted APOBEC-mediated mutagenesis, infrequent homologous recombination deficiency (HRD) signatures, and unique indel phenomena, such as the long-insertion signature in leukemia. Adult tumors, in contrast, manifest broader mutational diversity, are frequently shaped by exogenous exposures and therapeutic imprints, and exhibit elevated prevalence of APOBEC activity and HRD in specific tissue contexts. Children's cancers emerge under compressed mutational timelines, forcing a reliance on alternative pathways to malignancy, such as structural variants (including copy number variations, or CNVs) and epigenetic dysregulation, rather than the accumulation of single-nucleotide variants. Adults, by contrast, develop cancer through prolonged mutational time combined with a complex mixture of exposure-driven and age-related changes, reflecting both the passage of time and the cumulative impact of environmental insults and therapies.

An important limitation in detecting mutational signatures is the difference between whole-exome sequencing (WES) and whole-genome sequencing (WGS). While WES captures protein-coding regions and enables detection of dominant signatures, its restricted genomic coverage reduces sensitivity for low-burden or spatially diffuse mutational processes. In particular, it is important to note that signatures based on structural variants, including copy-number variants, will not be properly detected using exome data. In contrast, while genome sequencing provides a more comprehensive representation of the mutational landscape, it remains less accessible in pediatric oncology due to high cost and sample availability. These differences must be considered when comparing mutational signature detection and prevalence across studies.

Mutational Signatures in Hepatoblastoma

To illustrate these principles in a specific pediatric malignancy, hepatoblastoma, a cancer usually diagnosed within the first 3 years of life, is particularly instructive, as it combines an extremely low mutation burden with an incomplete molecular understanding. In the most recent study, which analyzed mutational signatures in 27 pediatric cancer subtypes, pilocytic astrocytoma had the lowest overall somatic mutation burden (median of 0.034 total mutations per megabase)⁹. However, notably, hepatoblastoma demonstrated an even lower mutation frequency (0.02 mutations per megabase) in a previous pan-cancer analysis⁷, albeit based on a smaller sample size of 16 tumors, highlighting potential inter-study variability due to cohort composition and sequencing depth. Hepatoblastoma is the most common liver malignancy in children, with an incidence of 2 per 1,000,000 children²⁰. Associated risk factors include congenital malformations, mutations in cancer predisposition genes, low birth weight, premature birth, and parental smoking^21-26. Despite its rarity and favorable 5-year survival rate (80%)²⁰, hepatoblastoma still presents a poorly understood etiology, as well as a scarcity of biomarkers for risk stratification, especially for high-risk cases.

In 2020, our group evaluated exome data from hepatoblastomas and identified three single-base substitution signatures, named HB-S1, HB-S2, and HB-S3²⁷. HB-S1 had similarities to COSMIC single-base substitutions S1 and S6, with distinct mechanistic origins: SBS1 is a clock-like, age-related signature associated with aging, while SBS6 is associated with defective DNA mismatch repair found in microsatellite-unstable tumors. The HB-S2 signature resembled COSMIC SBS29, a signature identified in cancers from individuals with a history of tobacco chewing. HB-S3 showed no correspondence with known signatures and exhibited an unspecific pattern of C>A transversions, suggesting a previously unrecognized mutational process specific to hepatoblastoma.

Although the association between parental smoking and hepatoblastoma risk has been previously described in the literature, subsequent studies did not corroborate this finding, possibly due to statistical issues associated with the rarity of hepatoblastomas^24,28,29. However, the presence of the tobacco-associated mutational signature SBS29 is not strictly restricted to the habit of smoking itself. As discussed in our prior study²⁷, SBS29 was originally revealed in gingiva–buccal oral squamous cell carcinomas arising in individuals with a tobacco-chewing habit. SBS29 may be associated with DNA damage caused by NNK, a tobacco-specific nitrosamine formed from nicotine metabolism; this damage includes the formation of O⁶- methylguanine DNA adducts, which are primarily repaired through transcription-coupled nucleotide excision repair and contribute to tobacco-related mutagenesis³⁰. Among the carcinogenic compounds present in smokeless tobacco, nitrosamines represent the most harmful class, with cancer risk directly correlating with exposure levels, as comprehensively reviewed by the International Agency for Research on Cancer³¹. Importantly, nitrosamines are not exclusive to tobacco products and are also present in dietary sources such as cured meats, smoked fish, and certain alcoholic beverages, including beer³². Supporting the biological relevance of early-life exposure, O⁶-methylguanine DNA adducts, hallmarks of nitrosamine-induced DNA damage, have been detected in human cord blood from mothers exposed to these compounds, indicating transplacental exposure and fetal toxicity³³. Consistent with this mechanism, maternal exposure to N-nitroso compounds or their precursors during pregnancy has been associated with adverse perinatal outcomes, including preterm birth³⁴, and with an increased risk of childhood cancer, specifically neuroblastoma and hepatoblastoma³⁵. Given that childhood cancers likely arise from the interplay between genetic susceptibility and early-life environmental exposures, we proposed that the converging lines of evidence support a model in which prenatal or early postnatal exposure to nitrosamines, originating from tobacco-related or dietary sources, may contribute to the emergence of SBS29-like mutational patterns in a subset of hepatoblastomas.

The PeCan platform from St. Jude Cloud [8] analyzed 11 hepatoblastoma samples and identified a signature profile with predominance of clock-like processes (SBS1 and SBS5), alongside ROS-associated damage signature (SBS18), aristolochic acid–associated signature (SBS22), platinum chemotherapy–related mutations (SBS31), and a signature of unknown etiology (SBS12). Nagae and colleagues (2021)¹⁴ examined 33 hepatoblastoma samples and similarly identified the predominant clock-like signatures (SBS1 and SBS5) and ROS-associated signature (SBS18). It is interesting to note that their analysis revealed no correlation between signature prevalence and either age at diagnosis or birth weight. This finding suggests that these mutational processes operate independently of these clinical variables, likely reflecting the intrinsic biological properties of hepatoblastomas rather than developmental or perinatal factors. Figure 2 summarizes the mutational signatures identified across hepatoblastoma cohorts, illustrating the predominance of SBS1, SBS5, and SBS18, as well as the enrichment of therapy-associated signatures, such as SBS31 and SBS35, in post-chemotherapy samples.

It is worth mentioning again that clock-like signatures (SBS1, SBS5), despite their nomenclature, do not reflect patient age but rather error-prone DNA replication during active cell division. In hepatoblastoma, these signatures are predominant, likely due to the high proliferative activity of hepatic progenitor cells during fetal and early childhood development, periods of intense liver growth and differentiation. However, we can hypothesize that the predominance of SBS1 in hepatoblastomas also reflects dysregulated epigenetic dynamics rather than solely replication errors. Notably, 5-hydroxymethylcytosine (5hmC), the oxidized derivative of 5-methylcytosine generated by TET family enzymes, was shown by us to be enriched in hepatoblastomas³⁶, associated with upregulated TET expression. The TET (Ten–Eleven Translocation) family of enzymes comprises TET1, TET2, and TET3, which catalyze the conversion of 5-methylcytosine to 5-hydroxymethylcytosine, playing a central role in active DNA demethylation and epigenetic regulation during development and differentiation. The TET-driven conversion of 5-methylcytosine to 5hmC, followed by spontaneous deamination and base excision repair, produces C>T transitions at CpG dinucleotides, the characteristic pattern of SBS1. This suggests that increased 5hmC flux and TET-mediated epigenetic remodeling drive recurrent deamination events in hepatoblastoma, providing a mechanistic explanation for the predominance of SBS1 in these tumors. This epigenetic–mutational coupling distinguishes hepatoblastoma from typical age-related SBS1 accumulation in adult cancers, instead linking its origin to dysregulated developmental epigenetic programs.

In one of the largest mutational signature analyses of pediatric liver tumors, Hirsch and colleagues (2021)¹⁵ studied 126 tumor samples (104 hepatoblastomas, 10 hepatocellular carcinomas, 7 fibrolamellar carcinomas, and 5 hepatocellular adenomas), performing whole-genome sequencing of 65 tumor samples and whole-exome sequencing of 57. This analysis detected four SBS signatures, two doublet base substitution signatures (DBS), and five indel signatures, including a novel DBS, together with the established COSMIC signatures (SBS1, SBS5, SBS18, and SBS35), as well as DBS5, ID1, ID2, ID3, ID5, and ID8. The SBS35 and DBS5 signatures, both associated with prior platinum chemotherapy (similar to SBS31), exhibited striking enrichment in primary hepatoblastoma sampled after neoadjuvant platinum chemotherapy, particularly within the molecularly defined “Liver Progenitor” subtype - characterized by high proliferative capacity, immune-cold phenotype, 11p15.5 alterations, and MDM4 amplifications. Further investigation revealed that liver progenitor cells could proliferate during chemotherapy exposure, accumulating SBS35 mutations after bypassing cisplatin DNA adducts. This finding established SBS35 as a marker of poor prognosis¹⁵. Subsequent validation by Pire and colleagues (2024)¹⁶ in an independent group of 54 hepatoblastomas resected after neoadjuvant chemotherapy reinforced this association: 33% of the samples exhibited this signature, and 56% of the SBS35-positive cases were classified as poor responders to neoadjuvant chemotherapy. Moreover, Thatikonda and colleagues (2023)⁹ observed that the ID3 signature, which is associated with tobacco smoking and was reported in hepatoblastomas¹⁵, can be correlated with SBS31 and SBS35. This finding suggests that DNA damage caused by chemotherapy or other mechanisms may generate similar mutational patterns related to tobacco smoking.

The ROS-associated signature (SBS18) is one major operative process in hepatoblastomas and is mechanistically linked to the metabolic properties of normal hepatic tissue. The liver is uniquely positioned as a high-energy-demand organ responsible for lipid metabolism, protein synthesis, and xenobiotic detoxification, functions requiring intense mitochondrial oxidative phosphorylation and cytochrome P450–mediated enzymatic activity^37,38. Both processes necessarily generate ROS as metabolic byproducts. In fetal and early childhood hepatic development, cellular proliferation and differentiation amplify these metabolic demands, creating an environment of elevated endogenous ROS flux. This inherent oxidative stress may be further amplified in hepatoblastoma by dysregulated antioxidant defense systems (SOD, catalase, and glutathione pathways), impaired mitochondrial quality control, or epigenetic dysregulation of genes involved in the oxidative stress response. Consistent with this model, SBS18 is most pronounced in hepatoblastomas compared to pediatric malignancies arising from tissues with lower basal metabolic demands and lower ROS production. The distinctive metabolic and oxidative context of hepatic tumorigenesis thus imprints a characteristic mutational signature upon hepatoblastoma genomes. These metabolic features are consistent with our metabolomic and lipidomic observations, which indicate altered oxidative and mitochondrial pathways in hepatoblastoma¹⁷, further supporting the role of ROS-driven mutational processes in this tumor type.

This review underscores the pressing need for harmonized, multi-institutional genomic analyses to facilitate the robust identification and interpretation of mutational signatures and to accelerate their translation into clinically actionable biomarkers for pediatric oncology.

Conclusion

Altogether, current data suggest that SBS35 could be interpreted as a predictive biomarker of exposure and response to platinum-based chemotherapy, reflecting therapy-induced mutagenesis and the emergence of treatment-related resistance mechanisms, offering potential clinical utility for treatment stratification and optimization of treatment management. Despite these advances, research on hepatoblastoma mutational signatures remains constrained by significant limitations. Across multiple independent studies, hepatoblastoma cohorts have been relatively small and heterogeneous, ranging from 10 exome-sequenced samples²⁷ to 65 whole-genome samples in a mixed liver tumor cohort¹⁵, and derived from limited geographic and ethnic populations, raising questions about the generalizability and reproducibility of identified signatures. In addition, the clinical translation of SBS35 looks promising. While the feasibility of detecting SBS35 using low-pass whole-genome sequencing remains unexplored, such an approach could substantially improve accessibility and patient risk management if validated. Furthermore, the use of SBS35 as a biomarker of chemoresistance depends on the availability of post-chemotherapy tumor material. Finally, its incorporation into current risk-stratification frameworks will require validation in larger, geographically diverse hepatoblastoma cohorts.

These limitations notwithstanding, the emerging picture of hepatoblastoma's mutational landscape reveals several insights: clock-like signatures predominate, reflecting endogenous replicative processes rather than exogenous exposures; signatures such as SBS35 correlate with chemotherapy resistance and poor prognosis; and ROS-associated and disease-specific mechanisms remain partially characterized. Similar challenges, including small and geographically limited cohorts, unknown etiologies, and scarce biomarkers, plague many other rare pediatric malignancies. Expanded, multi-institutional, and globally representative studies, such as the Pediatric Hepatic International Tumor Trial (PHITT)³⁹, are essential to overcome these obstacles in hepatoblastoma research.

Acknowledgments

This work was supported by the Brazilian funding agencies CNPq (401564/2025-8) and FAPESP (2023/17465-8).

Figure 1: Developmental timing of mutational processes in adult versus pediatric cancers.The scheme compares how mutational processes accumulate DNA lesions across the lifetime in adulthood tumors (left side of the figure) or during development in pediatric cancers (right side). Adult tumors acquire mutations gradually over decades, driven by signatures such as SBS1 (clock-like), SBS4 (tobacco-related), and SBS7a (UV-related), reflecting long-term environmental exposures and aging. In contrast, pediatric tumors develop within a compressed timeline and are dominated by endogenous processes, particularly the clock-like signatures SBS1 and SBS5, as well as the ROS-associated SBS18, consistent with limited exposure to external mutagens. The colored bars next to each signature indicate the relative contributions of different base-substitution types (e.g., C>T, C>A, T>G), illustrating the distinct genomic imprints left by each mutational process. Additionally, the colored circles, stars, and triangles within the cells represent different classes of mutations: circles for passenger mutations, stars for driver mutations, and triangles for chemotherapy resistance, emphasizing how tumors progressively accumulate diverse DNA alterations over time.

Figure 2: Mutational signature landscape of hepatoblastoma across diverse cohorts. This figure summarizes the mutational signatures identified across hepatoblastoma sequencing studies. Each circle size indicates the prevalence of a specific signature within each cohort (high prevalence, medium prevalence, and signatures that were absent or detected at very low levels). Empty squares correspond to signatures that were not evaluated in that particular study. Across datasets, the most recurrent COSMIC-aligned signatures are SBS1, SBS5, SBS18, and SBS35 (detected in part of tumors exposed to platinum-based chemotherapy). In our 2020 study (Aguiar et al., 2020, n=10 HB, WES), we identified three hepatoblastoma-specific mutational signatures: HB-S1, HB-S2, and HB-S3. HB-S1 shared features with COSMIC SBS1 and SBS6, reflecting a combination of age-related mutagenesis (SBS1) and mismatch-repair defects (SBS6). HB-S2 resembled SBS29, a signature associated with tobacco-chewing exposure, whereas HB-S3 showed no match to any known COSMIC signature and was characterized by an unspecific pattern of C>A transversions, suggesting a novel, previously unrecognized mutational mechanism in hepatoblastoma.

References

1. Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 2009;458:719–24. https://doi.org/10.1038/nature07943.
2. Helleday T, Eshtad S, Nik-Zainal S. Mechanisms underlying mutational signatures in human cancers. Nat Rev Genet 2014;15:585–98. https://doi.org/10.1038/nrg3729.
3. Kucab JE, Zou X, Morganella S, Joel M, Nanda AS, Nagy E, et al. A Compendium of Mutational Signatures of Environmental Agents. Cell 2019;177:821-836.e16. https://doi.org/10.1016/j.cell.2019.03.001.
4. Alexandrov LB, Kim J, Haradhvala NJ, Huang MN, Tian Ng AW, Wu Y, et al. The repertoire of mutational signatures in human cancer. Nature 2020;578:94–101. https://doi.org/10.1038/s41586-020-1943-3.
5. Yang X, Wang Q, Sun Y, Zhang Z, Wuchty S, Liang Z, et al. Genomic differences of patients with hematologic malignancies in different age groups. Commun Biol 2024;7:1630. https://doi.org/10.1038/s42003-024-07293-0.
6. Chen X, Yang W, Roberts CWM, Zhang J. Developmental origins shape the paediatric cancer genome. Nat Rev Cancer 2024;24:382–98. https://doi.org/10.1038/s41568-024-00684-9.
7. Gröbner SN, Worst BC, Weischenfeldt J, Buchhalter I, Kleinheinz K, Rudneva VA, et al. The landscape of genomic alterations across childhood cancers. Nature 2018;555:321–7. https://doi.org/10.1038/nature25480.
8. McLeod C, Gout AM, Zhou X, Thrasher A, Rahbarinia D, Brady SW, et al. St. Jude cloud: A pediatric cancer genomic data-sharing ecosystem. Cancer Discov 2021;11:1082–99. https://doi.org/10.1158/2159-8290.CD-20-1230.
9. Thatikonda V, Islam SMA, Autry RJ, Jones BC, Gröbner SN, Warsow G, et al. Comprehensive analysis of mutational signatures reveals distinct patterns and molecular processes across 27 pediatric cancers. Nat Cancer 2023;4:276–89. https://doi.org/10.1038/s43018-022-00509-4.
10. Alexandrov LB, Jones PH, Wedge DC, Sale JE, Campbell PJ, Nik-Zainal S, et al. Clock-like mutational processes in human somatic cells. Nat Genet 2015;47:1402–7. https://doi.org/10.1038/ng.3441.
11. Moore L, Cagan A, Coorens THH, Neville MDC, Sanghvi R, Sanders MA, et al. The mutational landscape of human somatic and germline cells. Nature 2021;597:381–6. https://doi.org/10.1038/s41586-021-03822-7.
12. Spisak N, de Manuel M, Milligan W, Sella G, Przeworski M. The clock-like accumulation of germline and somatic mutations can arise from the interplay of DNA damage and repair. PLoS Biol 2024;22:e3002678. https://doi.org/10.1371/journal.pbio.3002678.
13. Petljak M, Alexandrov LB, Brammeld JS, Price S, Wedge DC, Grossmann S, et al. Characterizing Mutational Signatures in Human Cancer Cell Lines Reveals Episodic APOBEC Mutagenesis. Cell 2019;176:1282-1294.e20. https://doi.org/10.1016/j.cell.2019.02.012.
14. Nagae G, Yamamoto S, Fujita M, Fujita T, Nonaka A, Umeda T, et al. Genetic and epigenetic basis of hepatoblastoma diversity. Nat Commun 2021;12. https://doi.org/10.1038/s41467-021-25430-9.
15. Hirsch TZ, Pilet J, Morcrette G, Roehrig A, Monteiro BJE, Molina L, et al. Integrated genomic analysis identifies driver genes and cisplatin-resistant progenitor phenotype in pediatric liver cancer. Cancer Discov 2021;11:2524–43. https://doi.org/10.1158/2159-8290.CD-20-1809.
16. Pire A, Hirsch TZ, Morcrette G, Imbeaud S, Gupta B, Pilet J, et al. Mutational signature, cancer driver genes mutations and transcriptomic subgroups predict hepatoblastoma survival. Eur J Cancer 2024;200. https://doi.org/10.1016/j.ejca.2024.113583.
17. Tasic L, Avramović N, Jadranin M, Quintero M, Stanisic D, Martins LG, et al. High-Resolution Magic-Angle-Spinning NMR in Revealing Hepatoblastoma Hallmarks. Biomedicines 2022;10. https://doi.org/10.3390/biomedicines10123091.
18. Alexandrov LB, Stratton MR. Mutational signatures: the patterns of somatic mutations hidden in cancer genomes. Curr Opin Genet Dev 2014;24:52–60. https://doi.org/10.1016/j.gde.2013.11.014.
19. Bouvard V, Loomis D, Guyton KZ, Grosse Y, Ghissassi F El, Benbrahim-Tallaa L, et al. Carcinogenicity of consumption of red and processed meat. Lancet Oncol 2015;16:1599–600. https://doi.org/10.1016/S1470-2045(15)00444-1.
20. Kahla JA, Siegel DA, Dai S, Lupo PJ, Foster JH, Scheurer ME, et al. Incidence and 5-year survival of children and adolescents with hepatoblastoma in the United States. Pediatr Blood Cancer 2022;69. https://doi.org/10.1002/pbc.29763.
21. Aguiar T, Teixeira A, Scliar MO, Sobral de Barros J, Lemes RB, Souza S, et al. Unraveling the Genetic Architecture of Hepatoblastoma Risk: Birth Defects and Increased Burden of Germline Damaging Variants in Gastrointestinal/Renal Cancer Predisposition and DNA Repair Genes. Front Genet 2022;13. https://doi.org/10.3389/fgene.2022.858396.
22. Chen H, Guan Q, Guo H, Miao L, Zhuo Z. The Genetic Changes of Hepatoblastoma. Front Oncol 2021;11. https://doi.org/10.3389/fonc.2021.690641.
23. Sobel Naveh NS, Traxler EM, Duffy KA, Kalish JM. Molecular networks of hepatoblastoma predisposition and oncogenesis in Beckwith‐Wiedemann syndrome. Hepatol Commun 2022;6:2132–46. https://doi.org/10.1002/hep4.1972.
24. Spector LG, Birch J. The epidemiology of hepatoblastoma. Pediatr Blood Cancer 2012;59:776–9. https://doi.org/10.1002/pbc.24215.
25. Tan ZH, Lai A, Chen CK, Chang KTE, Tan AM. Association of trisomy 18 with hepatoblastoma and its implications. Eur J Pediatr 2014;173:1595–8. https://doi.org/10.1007/s00431-013-2147-8.
26. Trobaugh-Lotrario AD, López-Terrada D, Li P, Feusner JH. Hepatoblastoma in patients with molecularly proven familial adenomatous polyposis: Clinical characteristics and rationale for surveillance screening. Pediatr Blood Cancer 2018;65:e27103. https://doi.org/10.1002/pbc.27103.
27. Aguiar TFM, Rivas MP, Costa S, Maschietto M, Rodrigues T, Sobral de Barros J, et al. Insights Into the Somatic Mutation Burden of Hepatoblastomas From Brazilian Patients. Front Oncol 2020;10. https://doi.org/10.3389/fonc.2020.00556.
28. McLaughlin CC, Baptiste MS, Schymura MJ, Nasca PC, Zdeb MS. Maternal and Infant Birth Characteristics and Hepatoblastoma. Am J Epidemiol 2006;163:818–28. https://doi.org/10.1093/aje/kwj104.
29. Spector LG, Johnson KJ, Soler JT, Puumala SE. Perinatal risk factors for hepatoblastoma. Br J Cancer 2008;98:1570–3. https://doi.org/10.1038/sj.bjc.6604335.
30. Pegg AE. Repair of O6-alkylguanine by alkyltransferases. Mutation Research/Reviews in Mutation Research 2000;462:83–100. https://doi.org/10.1016/S1383-5742(00)00017-X.
31. Cogliano V, Straif K, Baan R, Grosse Y, Secretan B, Ghissassi F El. Smokeless tobacco and tobacco-related nitrosamines. Lancet Oncol 2004;5:708. https://doi.org/10.1016/S1470-2045(04)01633-X.
32. Park J, Seo J, Lee J, Kwon H. Distribution of Seven N-Nitrosamines in Food. Toxicol Res 2015;31:279–88. https://doi.org/10.5487/TR.2015.31.3.279.
33. Annola K, Heikkinen AT, Partanen H, Woodhouse H, Segerbäck D, Vähäkangas K. Transplacental Transfer of Nitrosodimethylamine in Perfused Human Placenta. Placenta 2009;30:277–83. https://doi.org/10.1016/j.placenta.2008.12.012.
34. Vuong AM, Shinde MU, Brender JD, Shipp EM, Huber JC, Sharkey JR, et al. Prenatal Exposure to Nitrosatable Drugs, Dietary Intake of Nitrites, and Preterm Birth. Am J Epidemiol 2016;183:634–42. https://doi.org/10.1093/aje/kwv250.
35. Deng C, Hu Y, Sirirungreung A, Chock EY, Liew Z, Ritz B, et al. Exposure to nitrosatable drugs during pregnancy and childhood cancer: A cohort study in Taiwan. Int J Cancer 2026;158:2092–101. https://doi.org/10.1002/ijc.70240.
36. Rivas MP, Aguiar TFM, Fernandes GR, Caires LC, Goulart E, Telles-Silva KA, et al. TET upregulation leads to 5-hydroxymethylation enrichment in hepatoblastoma. Front Genet 2019;10. https://doi.org/10.3389/fgene.2019.00553.
37. Cederbaum AI. Cytochrome P450 2E1‐dependent oxidant stress and upregulation of anti‐oxidant defense in liver cells. J Gastroenterol Hepatol 2006;21. https://doi.org/10.1111/j.1440-1746.2006.04595.x.
38. Rui L. Energy Metabolism in the Liver. Compr. Physiol., Wiley; 2014, p. 177–97. https://doi.org/10.1002/cphy.c130024.
39. Schmid I, O’Neill AF, Watanabe K, Aerts I, Branchereau S, Brock PR, et al. Global consensus from the PHITT consortium surrounding an interim chemotherapy guidance for the treatment of children with hepatoblastoma, hepatocellular neoplasm – not otherwise specified (HCN-NOS), hepatocellular carcinoma and fibrolamellar carcinoma after trial closure. EJC Paediatric Oncology 2025;5:100235. https://doi.org/10.1016/j.ejcped.2025.100235.