Genomic profiling of Mexican patients with B‐cell precursor acute lymphoblastic leukemia reveals clinically significant somatic and potential germline variants Authors: Daniel Martínez Anaya, María del Rocío Juárez‐Velázquez, Ulises Juárez Figueroa, Michael Dean, Consuelo Salas Labadía, Marian Valladares Coyotecatl, Adriana Reyes León, Norma López Santiago, Luis Juárez Villegas, Marta Zapata Tarrés, Patricia Pérez‐Vera Published: 2026 ABSTRACT AbstractB‐cell precursor acute lymphoblastic leukemia (preB‐ALL) is characterized by pathogenic variants currently used in precision oncology. However, the mutational landscape of Mexican children with preB‐ALL has not yet been thoroughly explored and defined in terms of the clinical significance. We used a custom‐designed next‐generation sequencing exome panel, along with high‐resolution chromosome microarrays and gene fusion‐targeted assays, to characterize the mutational landscape of 73 Mexican children with preB‐ALL, exploring the profile of clinically significant variants. We classified the variants following the AMP‐ASCO‐CAP 2017 and ACMG‐CG 2019 guidelines recommendations. The mutational landscape includes a broad molecular spectrum of variants affecting cell cycle regulation, B‐cell development, kinase signaling, and epigenetic regulation genes. Tier 1 diagnostic variants allowed the identification of pre‐B ALL genetic subtypes, diminishing the preB‐ALL NOS group from 63% to 37%. Furthermore, 37% of cases presented Tier 1 variants conferring intermediate to adverse prognoses (primarily involving CRLF2 gene fusions, as well as PAX5, IKZF1, and TP53 inactivating mutations). Notably, these patients exhibited high‐risk clinical features and a lower event free survival rate than patients without these variants (60% versus 41.2%; p = 0.046; 95% CI). Between 19% and 42% of patients had Tier 2 variants targetable with JAK‐STAT or RAS‐MAPK signaling inhibitors. These patients showed a lower overall survival rate than patients without these variants (64% versus 90%; p = 0.048; 95% CI). Tier 1 potential germline variants in cancer predisposition genes (mainly BRCA1/2 and CHEK2) were observed in 10% of patients, some of whom had a family history of cancer. This highlights the importance of genetic counseling for patients and their families. Finally, 33% of patients had Tier 3 variants that were predicted to be deleterious and potentially upgraded to pathogenic with plausible clinical relevance. In conclusion, the mutational landscape analysis revealed variants useful for oncologic management and genetic counseling of Mexican children with preB‐ALL.Keywords: B‐cell precursor acute lymphoblastic leukemia, mutational landscape, tier 1 variants, tier 2 variants, potential germline variants, variants of uncertain significance, Mexican children Introduction B‐cell precursor acute lymphoblastic leukemia (preB‐ALL) is the most common childhood malignancy worldwide and a significant public health issue in Mexico [1]. This hematological cancer is characterized by the malignant proliferation of B‐cell progenitors that are arrested at an early stage of cellular differentiation [2]. PreB‐ALL is driven by a constellation of pathogenic variants, most of which are somatic changes that coexist with chromosomal abnormalities and act as cooperative genetic lesions that can trigger leukemic transformation [2]. Pathogenic variants can serve as biomarkers that define specific preB‐ALL molecular subtypes and confer prognostic value or predict responses to specific therapeutic molecules [3]. In addition, high‐ or moderate‐penetrance germline pathogenic variants involving cancer predisposition syndrome genes are associated with a neoplastic spectrum which includes preB‐ALL manifestation [4]. While significant progress has been made in describing the genomic landscape of preB‐ALL globally, data specific to the Mexican pediatric population remain underexplored, as previous studies have only described the occurrence of classical pathogenic variants [5, 6, 7, 8, 9, 10, 11]. In this study, we used a custom‐designed next‐generation sequencing (NGS) exome panel alongside high‐resolution chromosome microarrays and gene fusion‐targeted assays to comprehensively characterize the mutational landscape of 73 Mexican children with preB‐ALL. Based on the currently available evidence, we evaluated the clinical significance of these variants in terms of their predictive value for diagnosis, prognosis, and therapeutic response. Additionally, we evaluated variants of uncertain significance to select those that may be potentially upgraded as pathogenic. We present a comprehensive overview of the genomic landscape, including somatic and potential germline variants, classified based on their qualities as strong or potential clinical biomarkers for the management of Mexican children with pre‐B ALL. Materials and methods Ethics statement The study was conducted in accordance with local legislation and institutional requirements. This study was approved by the Research and Ethics Committees of the participating institutions (project numbers 067/2014, 019/2016, and 032/2023), and the National Commission of Bioethics (registration number CONBIOETICA‐09‐CEI‐025‐20,161,215). Patient selection Seventy‐three patients aged ≤18 years were enrolled in this study. They were recruited consecutively between January 2015 and December 2018 at the National Institute of Pediatrics and the Children's Hospital of Mexico Federico Gómez, in Mexico City. Patients or their guardians signed informed assent/consent forms in accordance with the Declaration of Helsinki. Pre‐B ALL diagnosis was established through cytomorphology and flow cytometry immunophenotyping (CD10, CD19, CD20, CD34, and CD79) according to WHO 5th edition classification criteria for Hematolymphoid tumors, and the initial genetic classification is detailed in Supplementary Materials and Methods [12]. Biological samples and molecular assays Genomic DNA was extracted from bone marrow samples using a DNeasy Blood & Tissue kit (Qiagen, Düsseldorf, Germany). High‐quality DNA was subsequently used for custom‐designed NGS exome panel, chromosomal microarray analysis (CMA), or Sanger sequencing analysis. RNA was extracted using the RNeasy kit (Qiagen, Düsseldorf, Germany), and cDNA was obtained using standard methods (Invitrogen, Waltham, MA, USA) for RT‐qPCR assays. Finally, fixed nuclei were obtained for FISH testing. For NGS, exons of 205 pediatric cancer‐related genes were targeted and paired‐end sequencing was performed (supplementary material, Table S1). CMA was performed according to the manufacturer's instructions. Patients negative for sentinel gene fusions were evaluated for CRLF2::P2RY8/IGH gene fusions. The processing details of NGS, CMA, and CRLF2 gene fusion testing are provided in Supplementary Materials and Methods. Variant classification Synonymous changes, variants along exon boundaries or splicing regions, and variants with a variant allele frequency (VAF) <0.01 or a depth of coverage <10 were excluded. High‐quality variants were manually classified according to the four‐tier Association for Molecular Pathology/American Society of Clinical Oncology/College of American Pathologists 2017 criteria [3] using Franklin by Genoox version 80.2 (https://franklin.genoox.com/clinical‐db/home), CancerVar (https://cancervar.wglab.org), and OncoKB (https://www.oncokb.org) databases (accessed on 15 January 2026) [13, 14]. Polymorphic variants (MAF ≥1%) were identified based on population allele frequency values reported in the gnomAD Exomes American Admixed/Latinos, gnomAD Global (http://gnomad-sg.org), or dbSNP (https://www.ncbi.nlm.nih.gov/snp/) databases (accessed on 15 January 2026). Copy number variants (CNVs) with a minimum of five consecutive probes and copy number neutral loss of heterozygosity (CNN‐LOH) regions larger than 3 Mb and covered by at least 50 consecutive probes were analyzed if they included any of the 1,186 cancer genes recognized as commonly altered in cancer according to the OncoKB database (supplementary material, Table S2) [14]. Variants were manually classified according to the ACMG/CG 2019 guidelines [15]. CNN‐LOH regions were considered pathogenic when they involved genes commonly altered by inactivating mutations or epigenetic silencing in preB‐ALL or other hematological malignancies [15, 16, 17, 18]. The SNV/Indel variants were considered as potential germline if they involved a cancer predisposition gene (supplementary material, Table S3), identified in patients with an associated genetic syndrome, and/or exhibited a VAF of ~0.50 or 1.0 (or ≤ 0.30 if the genomic region was affected by a chromosomal imbalance) according to the guidelines proposed by Kraft and Godley, for the identification of potential germline alleles from NGS assay of tumor samples [19, 20]. Potential germline variants were reclassified according to the ACMG/AMP 2015 criteria using Franklin by Genoox, InterVar, and ClinVar databases [21]. Variant validation We used Sanger sequencing (SS) or FISH as appropriate. For SS, specific oligonucleotides were designed (supplementary material, Table S4), and the reactions were performed using bone marrow DNA following standard methods. FISH analyses were performed on fixed bone marrow cells. One hundred nuclei were analyzed using LSI and centromeric probes, in accordance with the manufacturer's recommendations. To validate the germline etiology, genomic DNA was obtained from saliva samples collected from patients harboring potential germline variants using a prepIT‐L2P kit (Genotek Inc., Ottawa, Canada). SS was performed on DNA from paired bone marrow and saliva samples. A variant was considered germline if it was detected in both tissues. Clinical significance assignment Following the AMP/ASCO/CAP 2017 and ACMG/CG 2019 criteria, we consulted the currently available evidence from OncoKB database, COG, NCCN, WHO 2022 professional ALL guidelines/international consensus, or specific clinical trials [12, 22, 23, 24]. The variants were classified as follows: Tier 1: Diagnostic, prognostic, and/or therapeutic clinical biomarkers for preB‐ALL patients, defining a specific genetic subtype, contributing to risk stratification, or qualifying (or predicting resistance) for FDA‐approved therapies. Tier 1: Diagnostic, prognostic, and/or therapeutic clinical biomarkers for preB‐ALL patients, defining a specific genetic subtype, contributing to risk stratification, or qualifying (or predicting resistance) for FDA‐approved therapies. Tier 2: Diagnostic, prognostic, and/or therapeutic clinical biomarkers for patients with different tumor types. Tier 2: Diagnostic, prognostic, and/or therapeutic clinical biomarkers for patients with different tumor types. Tier 3/ Variant of uncertain significance (VUS): Do not meet the criteria for Tiers 1/2, and there is no convincing published evidence of their clinical significance in preB‐ALL/cancer. However, they did not meet the criteria for the Tier 4 benign/likely benign classification. Tier 3/ Variant of uncertain significance (VUS): Do not meet the criteria for Tiers 1/2, and there is no convincing published evidence of their clinical significance in preB‐ALL/cancer. However, they did not meet the criteria for the Tier 4 benign/likely benign classification. Uncertain significance variants processing VUS derived from exome sequencing were further evaluated using bioinformatics prediction tools and functional context data obtained from public databases to identify variants with a potentially deleterious impact. VUS that fulfill the following five criteria were considered potentially pathogenic: (1) classified as deleterious by at least three different pathogenicity predictors, (2) affect a gene with a documented baseline expression in B‐cell precursors/leukemic cells, (3) locate within a protein domain, (4) decrease protein structural stability, and (5) involve a highly conserved residue. VUS that did not fulfill all these criteria were downgraded as potentially benign [25, 26, 27]. The bioinformatic tools and databases used for VUS processing are detailed in Supplementary Materials and Methods. Statistical analyses Categorical variables were compared using a two‐tailed Fisher´s exact test. Survival analyses were conducted in patients with complete clinical follow‐up data who adhered to the chemotherapy regimen between diagnosis and first relapse. Overall survival (OS), relapse‐free survival (RFS), and event‐free survival (EFS) curves were estimated using the Kaplan–Meier method and compared using the log‐rank test according to the presence/absence of Tier 1 variants associated with intermediate‐to‐adverse prognosis or Tier 2 targetable variants. All statistical analyses were performed using SPSS version 31.0.0.0 (IBM Corp., Armonk, NY, USA), and significant differences were defined as a two‐sided p‐value ≤0.05. Results Study cohort genetic profiling and variant classification A cohort of 73 Mexican pediatric patients diagnosed with preB‐ALL was analyzed. The 97% of the study cohort was enrolled at diagnosis; based on their genetic features, most of the patients were initially classified as not otherwise specified (preB‐ALL–NOS). The main clinical features of the study cohort are summarized in supplementary material, Table S5. All patients were analyzed using an NGS exome panel, 53 patients (73%) were complemented with CMA, and the entire cohort was tested for sentinel gene fusions; 29% were positive. The remaining patients were negative for sentinel gene fusions; therefore, they were eligible for CRLF2 gene fusion analysis, 84% were available for testing. CRLF2 overexpression was detected in 60% of the tested patients; CRLF2::P2RY8/IGH fusions were detected in 47% of patients positive for CRLF2 overexpression, while the remaining patients were negative or did not have available material for gene fusion testing. The genetic profile is summarized in supplementary material, Table S6. The average NGS coverage and CMA resolution were 376,91X and 750 K, respectively. This allowed for the accurate determination of the SNV/indel VAF and molecular breakpoints of the CNV/CNN‐LOH regions. The variant filtering process and classification are shown in supplementary material, Figure S1. A total of 1,835 high‐quality SNVs/Indels were identified in 205 genes sequenced from 73 preB‐ALL patients. The results were as follows: 6% of the total variants were identified as pathogenic [including Tier 1 (0.5%), potential germline pathogenic/likely pathogenic (0.5%), and Tier 2 variants (3%), and Tier 3 variants reclassified as potentially upgradable as pathogenic based on their predicted deleterious impact (2%)]; the remaining variants corresponded to Tier 3 potentially down‐gradable as benign, and Tier 4 variants. We identified 2,840 high‐quality CNV/CNN‐LOH variants encompassing OncoKB cancer genes in 53 patients; most of these variants were identified as Tier 3 or Tier 4, and only 1% and 4% corresponded to Tier 1 and Tier 2 variants, respectively. All gene fusions detected by targeted assays were classified as Tier 1 variants. Mutational landscape characterization Through the integration of genomic methodologies and gene fusion‐targeted assays, a mutational landscape was obtained, comprising Tier 1, Tier 2, and potentially pathogenic Tier 3/VUS variants (Figure 1). Notably, most variants exhibited a VAF between 0.25 and 0.65 or mosaic Log2R mean values, suggesting that these variants were somatic mutations that constituted the main clones. This hypothesis was corroborated through orthogonal validation. (supplementary material, Figures S2–S4). The mutational landscape involves genes associated with biological processes commonly deregulated in preB‐ALL. In general, cell cycle regulation, B‐cell development, kinase signaling, and DNA repair genes were mainly affected by Tier 1 (somatic and potential germline) and Tier 2 variants, whereas epigenetic regulation genes were affected by Tier 2 variants. Molecular landscape of Tier 1, Tier 2, and Tier 3 variants potentially upgradable as pathogenic. The oncoplot diagram shows the variants distribution in 73 Mexican children with pre‐B‐ALL. Each patient is represented by a column and is grouped according to the genetic subtype. Each row represents a gene, which is grouped based on biological processes. The upper graphic shows the number of variants per patient. The left graphic shows the variant frequency per gene (the tier classification is labeled with Arabic numbers). The bottom panel shows the coverage of the different genomic or targeted analyses performed on the study cohort. CMA, chromosome microarray analysis. Cell cycle regulation genes were affected in approximately 40% of patients, with Tier 2 monoallelic deletions involving the CDKN2A/B genes being the most common mutation, followed by Tier 1 variants affecting the TP53 gene. Variants affecting B‐cell development genes were observed in at least 55% of the patients. The most common mutations included Tier 2 deletions affecting the ETV6 gene and the IKZF1 and PAX5 Tier 1 inactivating variants (Figure 1). Variants affecting kinase signaling genes were the most prevalent (73%), principally Tier 1 fusions involving the CRLF2 gene and Tier 2 variants involving the RAS‐MAPK signaling genes. At least 41% of patients showed variants affecting epigenetic regulation genes. The mutational spectrum principally includes Tier 2 CNN‐LOH variants involving the ASXL1, EP300, and KMT2D genes. Finally, several genes associated with other biological processes were affected by Tier 3 variants potentially upgradable as pathogenic (Figure 1). Relating to clinical significance categories, 96% of the study cohort presented at least one variant from Tier 1, Tier 2, or Tier 3 potentially upgradable as pathogenic. Tier 1 variants were observed in 70% of the study cohort (with an average of 0.89 variants per patient), most of which had diagnostic and prognostic significance (Figure 2A). Tier 2 variants were observed in 77% of patients (with an average of 1.93 variants per patient), most of which had diagnostic or therapeutic significance (Figure 3A). Additionally, 10% of the study cohort showed Tier 1 potential germline variants in cancer‐predisposition genes (Figure 4A). Finally, Tier 3/VUS potentially upgradable as pathogenic were observed in 33% of patients (with an average of 0.41 variants per patient). Clinical impact of Tier 1 variants. (A) Frequency and clinical impact categories (B) Genetic and prognostic classification of patients before (inner circle) and after (outer circle) the Tier 1 variant clinical impact interpretation (C) Kaplan–Meier analyses of the study cohort based on the presence or absence of Tier 1 prognostic variants. Clinical impact of Tier 2 variants. (A) Frequency and clinical impact categories. (B) Frequency of potentially diagnostic and/or prognostic variants per gene. (C) Frequency of potentially therapeutic variants according to the genetic subtype. (D) Kaplan–Meier analyses of the study cohort based on the presence/absence of Tier 2 targetable variants. Potential germline Tier 1 variants involving cancer‐predisposing genes. (A) Frequency of potential germline Tier 1 variants. Created with BioRender.com. (B) Schematic representation of the localization of potential Tier 1 germline variants within the domains of proteins encoded by cancer‐predisposing genes. (C) Representative results of germline etiology validation by Sanger sequencing on paired saliva and bone marrow DNA. The clinical significance of each variant constituting the mutational landscape is detailed in supplementary material, Tables S7–S10 and the clinical impact on the study cohort is summarized in the following sections. Impact on diagnosis and prognosis Tier 1A diagnostic variants enabled the identification of seven preB‐ALL genetic subtypes, reducing the preB‐ALL–NOS group from 63% to 37% (Figure 2B). These subtypes include five classical and the following two new/provisional entities: PreB‐ALL–PAX5 P80R positive and pre‐B ALL with CRLF2 and/or PAX5 rearrangements, which are defined by gene fusions which promote CRLF2 overexpression and structural variants which disrupt the PAX5 domains (supplementary material, Figure S5A). Most of the Tier 1 diagnostic variants also possessed prognostic significance; 33% of the study cohort presented with Tier 1A favorable‐prognosis variants, including ETV6::RUNX1 and TCF3::PBX1 gene fusions and hyperdiploidy. Most of these patients were classified as NCI standard risk, had white blood cell counts <50,000 mm3, did not experience relapse, and were under surveillance (Table 1). Table 1Frequency of variants in the study cohort grouped by genetic subtypes and clinical featuresPatient groups, n (%)Tier 1Tier 1 potential germlineTier 2Tier 1A favorable prognosis † Tier 1A/B intermediate‐adverse prognosis § Tier 2 targetable variants ‡ Tier 3 (potentially pathogenic)Total cohort, 73 (100%)70%10%77%33%37%57%33%Genetic subtypes BCR::ABL1, 3 (4%)100%0%66.6%0%100%0%0% ETV6::RUNX1, 4 (5%)100%0%25%100%0%0%25% TCF3::PBX1, 7 (10%)100%0%57%100%0%28%28.5% CRLF2 and/or PAX5 rearrangement, 14 (19%)100%7%100%0%100%93%36% PAX5 P80R, 2 (3%)100%50%100%0%100%50%50%Hyperdiploidy, 13 (18%)100%31%85%100%3%91%54%iAMP21, 3 (4%)100%0%100%0%100%66.6%33.3%preB‐ALL NOS, 27 (37%)18.5%4%85%0%7%48%33.3%Clinical groupsHigh risk, 38 (52%)74%5.2%87%21%60%60%61%Standard risk, 33 (45%)66%12%64%48%21%51%36% p‐value0.600.400.028* 0.023* 0.0015* 0.480.057** WBCc ≥50,000 mm3, 22 (30%)77%9%95%23%54%54%41%WBCc <50,000 mm3, 48 (66%)69%10%77%37%35%52%33% p‐value0.571.00.080.270.191.00.59Positive extramedullary infiltration, 11 (15%)91%9%100%27%54%64%36%Negative extramedullary infiltration, 41 (56%)65.8%5%80%34%39%46%41% p‐value0.140.510.171.00.490.491.0Relapse, 17 (23%)70%6%70%11%59%53%47%No relapse, 32 (44%)72%6%75%53%28%28%44% p‐value1.001.000.740.0058* 0.060.121.00Death, 14 (19%)78%7%78%28%57%50%57%Survivors, 38 (52%)68%5%74%39%34%34%40% p‐value0.731.01.00.190.310.340.34WBCc, white blood cell counts.*Statistically significant.**Tendency towards statistical significance. † These variants include ETV6::RUNX1, TCF3::PBX1, and hyperdiploidy. ‡ These include activating mutations in RAS‐MAPK and JAK–STAT signaling genes. § These variants include BCR::ABL1, CRLF2, and/or PAX5 rearrangements, iAMP21, IKZF1, and TP53 inactivating variants. Frequency of variants in the study cohort grouped by genetic subtypes and clinical features WBCc, white blood cell counts. Statistically significant. Tendency towards statistical significance. These variants include ETV6::RUNX1, TCF3::PBX1, and hyperdiploidy. These include activating mutations in RAS‐MAPK and JAK–STAT signaling genes. These variants include BCR::ABL1, CRLF2, and/or PAX5 rearrangements, iAMP21, IKZF1, and TP53 inactivating variants. Thirty‐seven percent of the patients presented with Tier 1A/B variants conferring an intermediate‐to‐adverse prognosis, including BCR::ABL1 fusion, CRLF2 and PAX5 rearrangements, PAX5 P80R mutation, iAMP21 and TP53/IKZF1 inactivating variants, which are not diagnostic but were classified as Tier 1B variants associated with adverse prognostic significance. These variants are primarily monoallelic deletions; however, TP53 is also affected by SNV/indels clustered in the DNA‐binding domain (supplementary material, Figure S5A). Most patients harboring Tier 1 variants conferring intermediate‐to‐adverse prognosis were classified as NCI high‐risk (p = 0.0015), exhibited clinical features associated with an unfavorable outcome, and showed a lower EFS rate than patients without these variants (60% versus 41.2%; p = 0.046; 95% CI). Lower overall and RFS rates were also observed, approaching statistical significance (Table 1; Figure 2C). Finally, 77% of the study cohort presented Tier 2 variants that help with the diagnosis of cancer, as they involve genes recurrently mutated in hematological and solid tumors (such as CDKN2A/B, NRAS/KRAS, and ETV6). Between 1% and 12% presented Tier 2 diagnostic variants with prognostic significance in different types of hematological neoplasms based on average quality evidence. These are inactivating variants involving mainly ASXL1, SH2B3, and PTPN11 (Figure 3B). Impact on therapy BCR::ABL1 was the only actionable variant with strong therapeutic significance and was observed in 4% of patients. However, 57% of the patients presented Tier 2 targetable variants identified as having therapeutic potential. The most prevalent mutations affected genes associated with the RAS‐MAPK signaling pathway (42%) and were observed mainly within the hyperdiploid and preB‐ALL–NOS groups (Figure 3C). These variants mostly affect oncogenes and negative regulators of the RAS‐MAPK signaling cascade. The main mutational event is missense SNVs involving the hotspot residues Gly12/13 in the catalytic domain of NRAS/KRAS (supplementary material, Figure S5B). The JAK‐STAT activating mutations were observed in ~19% of the patients, mainly within the CRLF2 and/or PAX5 rearranged subtype. These variants included the CRLF2 gene fusions (classified as Tier 1 diagnostic/prognostic but downgraded to Tier 2 for therapeutic significance), the JAK2p.R683G mutation, and the amplification of the 5' untranslated region of the FGFR2 gene, observed in 7% of patients (Figure 3C; supplementary material, Figure S5B). Interestingly, Tier 2 actionable variants were more frequent in patients negative for sentinel gene fusions, who relapsed or died by disease progression (Table 1). These patients showed a lower OS than patients negative for these variants (64% versus 90%; p = 0.048; 95% CI) (Figure 3D). Five patients (7%) presented variants conferring resistance to standard chemotherapy: (1) The Tier 1B TP53p.R248Q mutation (3%), observed in one patient at diagnosis (ALL–120), and one at relapse (ALL–020). (2) The Tier 2D CREBBP inactivating variants, which disrupt the histone acetyltransferase domain, were observed in three patients at diagnosis (4%) (Figure 3C; supplementary material, Figure S5B). Potential germline variants in cancer predisposition genes Carriers of Tier 1A potential germline variants showed missense, nonsense, and frameshift mutations involving DNA repair and RAS signaling negative regulator genes associated with cancer predisposition (Figure 4A,B). Germline etiology was confirmed in 30% of cases (Figure 4C), including one patient diagnosed with type 1 neurofibromatosis and two patients with a family history of cancer. The germline etiology could not be validated in the remaining 70% of patients because they did not provide consent. However, these patients showed variants that were previously recorded in ClinVar as pathogenic or likely pathogenic germline variants (supplementary material, Table S8). Regarding preB‐ALL features, these patients were observed mainly within the hyperdiploid cases. Almost half of these patients did not relapse and are currently under surveillance (Table 1). Tier 3/VUS variants potentially upgradable as pathogenic Eight percent of the Tier 3/VUS variants identified through exome sequencing were reclassified as potentially pathogenic because of their predicted deleterious effects (Figure 5A; supplementary material, Table S10). These were missense variants observed in 33% of the analyzed patients, most of whom were negative for sentinel gene fusions or presented with high‐risk clinical features (Table 1). Most variants were non‐recurrent, and all of them involved regulatory or catalytic domains of genes responsive to kinase inhibitors, were associated with chemoresistance, or were recognized as cancer‐predisposing genes. Considering these variants, the frequency of patients with mutations with a potential therapeutic impact increased from 58% to 60%, and the frequency of carriers of potential germline variants with a deleterious effect increased from 10% to 15% (Figure 5B,C; supplementary material, Table S10). Tier 3 variants potentially upgradeable as pathogenic. (A) Flowchart of the VUS processing. (B) Frequency of Tier 3/VUS variants potentially upgradable to pathogenic variants per gene. (C) Representative examples of potentially deleterious Tier 3/VUS variants. Schematic representations of protein diagrams showing the localization of variants within protein domains and screenshots of the results of the conservative analyses and pathogenicity predictions. Discussion This study analyzed the mutational landscape of 73 Mexican children with preB‐ALL and classified the variants based on their clinical significance. Molecular landscape insights We identified a heterogeneous spectrum of genomic alterations associated with pediatric preB‐ALL. Some genes (CDKN2A/B and NRAS/KRAS) were hit by recurrent mutations involving hotspot regions, whereas others (such as TP53, PAX5, and SH2B3) exhibited a broad mutational spectrum. This molecular heterogeneity constitutes a technical challenge, suggesting that using a single genetic screening methodology may lead to an underestimation of the mutational landscape, prompting the use of multiple genetic screening techniques to maximize the detection of the genetic variants [28, 29, 30]. Impact on clinical management of preB‐ALL The mutational landscape analysis had a significant diagnostic impact, reducing the preB ALL–NOS group from 63% to 37%. Most Tier 1A/B diagnostic variants corresponded to CNVs and gene fusions, highlighting the relevance of structural alterations in this population. Similarly, Sánchez‐Escobar et al characterized a cohort of Mexican pediatric patients with preB‐ALL using RNA‐seq and identified all the genetic subtypes, mainly driven by Tier 1 CNVs and gene fusions [31]. These findings suggest that a combined CMA/RNA‐seq strategy could resolve the preB‐ALL–NOS group in Mexican pediatric patients with preB‐ALL by improving the detection of structural variants not fully captured by targeted exome panels. Applying a CMA/RNA‐seq combined strategy also may lead to the reclassification of some cases within the CRLF2 and/or PAX5 rearranged subtype as Ph‐like‐preB‐ALL, considering the gene expression signature overlapping between these molecular entities [24, 32]. Consistent with this hypothesis, some of these patients exhibited a Ph‐like mutational background with IKZF1 deletions and activating mutations in tyrosine kinase receptor genes, such as FGFR2. The Ph‐like subtype has been identified as highly frequent in Mexican pediatric patients, reinforcing this possibility [33]. The frequency of preB‐ALL genetic subtypes can be influenced by age, ancestry, and analytical methods. The frequencies of BCR::ABL1 fusion and PAX5 rearrangements were consistent with those expected. The frequency of hyperdiploidy could be higher, as some NOS patients could not be analyzed by CMA. We observed a lower frequency of ETV6::RUNX1 in our study cohort than in Hispanics from the US and a higher frequency of iAMP21 and TCF3::PBX1 in Mexicans than in other populations [4, 11, 31, 33, 34, 35, 36]. Additionally, a higher frequency of the PAX5 P80R mutation was observed in our study cohort and a frequency of CRLF2::IGH rearrangement was similar to that reported in Hispanic patients from the United States and higher than non‐Hispanic patients (supplementary material, Table S11) [37]. Regarding prognosis impact, 63% patients presented Tier 1A diagnostic variants, which also conferred a specific prognosis. However, the inactivating variants in IKZF1 and TP53 expanded the prognostic yield of the genomic screening analysis from 63% to 70%. This information allowed for more precise risk stratification for the preB‐ALL NOS and hyperdiploid patients who were positive for these underlying variants. This is clinically relevant, as these variants are associated with poor prognosis in childhood ALL, as supported by high‐quality evidence [38, 39, 40, 41]. The Tier 2 targetable variants are currently under investigation in ongoing clinical trials for pediatric leukemias. They are actionable with different FDA‐approved drugs for different types of cancer or confer resistance to conventional chemotherapeutic agents based on average‐quality evidence [42, 43]. Interestingly, these variants were particularly frequent among patients exhibiting features associated with poor outcomes, suggesting a potential benefit of applying targeted therapy. According to our results, 42% of the study cohort showed an RAS‐MAPK mutation and could potentially benefit from MEK inhibitor therapy including selumetinib or trametinib. The combination of these molecules with glucocorticoids has shown promising results in reducing the leukemic burden in mice engrafted with primary‐derived RAS‐mutated leukemia cells [42, 43]. An international phase I/II clinical trial tested the efficacy of the selumetinib and dexamethasone combination for treating adults and children with multiple relapses of refractory ALL. Preliminary results were promising, with 45% of patients achieving complete remission [44]. A JAK‐STAT‐targetable genetic background was detected in ~19% of the cohort, suggesting that these patients may benefit from therapy with ruxolitinib, a type‐I JAK1/2 inhibitor approved by the FDA for treating myeloproliferative neoplasms. The effectiveness of ruxolitinib against JAK‐mutated preB‐ALL has been demonstrated in preclinical models and case reports [45]. Phase II clinical trials are currently underway to determine the safety and efficacy of ruxolitinib in combination with standard chemotherapy for adolescents and adults with JAK‐deregulated ALL (supplementary material, Table S9). Although ruxolitinib is advancing in ongoing clinical trials, concerns remain because preclinical studies and case reports have shown reduced efficacy associated with the emergence of resistance mechanisms [45]. We detected resistance variants in patients recruited at the time of diagnosis and at the relapse, suggesting a concerning possibility of chemotherapy failure in some patients. While the role of TP53p.R248Q in reducing the effectiveness of DNA‐damaging drugs is supported by high‐quality studies, the chemoresistance associated with CREBBP inactivating variants remains controversial because average quality evidence supports their role in the reduction of response to glucocorticoids; however, other investigations suggest that CREBBP does not modulate the chemotherapeutic response. Therefore, the clinical relevance of these variants must be clarified [46, 47, 48, 49]. Cancer predisposition implications In different preB‐ALL pediatric cohorts, the frequency of germline pathogenic variants involving cancer predisposition genes has been documented at 8.3–11%, which is comparable to the 10% observed in this study [50, 51]. Some of these patients manifested associated genetic syndromes or had a family history of cancer. Therefore, genetic counseling and personalized oncological management are mandatory to avoid the potential side effects of conventional therapeutic strategies [4]. Insights from the uncertain significance variant analysis We identified Tier 3/VUS variants with a potentially deleterious effect based on in silico predictions and functional context data. It is important to note that this is a conservative estimation of pathogenicity, as some variants require further experimental validation. Nevertheless, based on new preclinical evidence reclassifying some of these variants as Tier 2 merits consideration. For instance, WHSC1 p.E1009K variant, observed in 3% of patients, has been characterized as a somatic loss‐of‐function variant that promotes epigenetic deregulation, tumor growth, and context‐dependent resistance to purine analogs [52, 53]. FLT3 and JAK1/2 mutations, observed in 3–4% of patients, have been characterized as gain‐of‐function mutations in preclinical models of acute leukemia and respond to inhibitor molecules, suggesting potential therapeutic value [54, 55, 56, 57]. Lastly, 1–3% of patients show potentially pathogenic VUS involving cancer predisposition genes; in these patients, the possibility of non‐sporadic preB‐ALL should be thoroughly investigated [58, 59]. Strengths and limitations This study represents the first comprehensive characterization of the clinically annotated molecular landscape of Mexican preB‐ALL patients. The integration of genomic and targeted methodologies strengthened variant detection across multiple genomic alteration types. Moreover, the identification of somatic and germline variants provides direct implications for oncological management and genetic counseling. However, some limitations should be acknowledged. The relatively small sample size may limit the population representativeness of variant frequencies. Additionally, not all patients underwent complete genomic screening, which could lead to an underestimation of the detection of certain variants, particularly non‐coding regions and structural variants. Despite the limitations, this study provides valuable insights into the mutational architecture of Mexican pediatric preB‐ALL and supports the implementation of integrated genomic strategies in this population. Author contributions statement DMA and PP‐V conceived and designed the study. DMA and UJF curated the genomic data. Formal analysis was performed by DMA, MdRJ‐V, MVC and CSL. Funding for the study was acquired by PP‐V. The investigation was conducted by DMA and PP‐V. Methodology was developed by DMA, MdRJ‐V and MD. Project administration and resource management were carried out by PP‐V. Software development was undertaken by DMA and UJF. Supervision was provided by NLS, LJV, MZT and ARL. The original draft of the manuscript was written by DMA and PP‐V. All authors, including DMA, MdRJ‐V, UJF, MD, CSL, MVC, ARL, NLS, LJV, MZT and PP‐V, contributed to the review and editing of the manuscript. All authors have read and agreed to the published version of this manuscript. Supporting information Supplementary materials and methods. Figure S1. Experimental study design, variant classification and categorization Figure S2. The variant allele fractions of Tier 1, Tier 1 potential germline, Tier 2, and Tier 3 potentially upgradeable as pathogenic Figure S3. Mean Log2ratio of Tier 1 and Tier 2 CNVs Figure S4. Representative results of variant orthogonal validation Figure S5. Schematic representation of the localization of representative Tier 1 and Tier 2 somatic variants Table S1. Genes screened by exon sequencing panel Table S2. Cancer genes registered in OncoKB database Table S3. Genes assessed by exome sequencing panel and recognized as cancer predisposition genes Table S4. Oligonucleotide sequences Table S5. Clinical features of the study cohort Table S6. Genetic profile of the study cohort Table S7. Clinical significance of SNV/Indels classified as Tier 1 or Tier 2 variants Table S8. Main features of Tier 1 potential germline variants Table S9. Clinical significance of CNVs, CNN‐LOH, and gene fusions classified as Tier 1 or Tier 2 Table S10. Tier 3 variants predicted as deleterious and reclassified as potentially pathogenic Table S11. Frequencies of genetic subtypes detected in this study and comparison with other cohorts Acknowledgements This paper is part of the requirements for obtaining a Doctoral degree at Posgrado en Ciencias Biológicas, UNAM, for author Daniel Martínez Anaya. We gratefully acknowledge the Secretaría de Ciencias, Humanidades, Tecnología e Innovación SECIHTI for the support to this research through a graduate scholarship (CVU: 816116). This work was supported in part by the National Council for Science and Technology (CONACYT) through grant PDCPN‐2004/248591, as well as by the National Institute of Pediatrics through grants 067/2014, 019/2016, and 032/2023, and by the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. The authors would like to thank Victor Cruz Martínez, Manuel Ruíz González, and Dafné Linda Moreno Lorenzana for their technical support. No conflicts of interest were declared. Data availability statement The main data supporting the results of this study are available in the paper and its supplementary information. All data generated and datasets analyzed in the current study are available from the corresponding author upon reasonable request. References References 60 – 67 are cited only in the supplementary material. Associated Data Supplementary Materials Supplementary materials and methods. Figure S1. Experimental study design, variant classification and categorization Figure S2. The variant allele fractions of Tier 1, Tier 1 potential germline, Tier 2, and Tier 3 potentially upgradeable as pathogenic Figure S3. Mean Log2ratio of Tier 1 and Tier 2 CNVs Figure S4. Representative results of variant orthogonal validation Figure S5. Schematic representation of the localization of representative Tier 1 and Tier 2 somatic variants Table S1. Genes screened by exon sequencing panel Table S2. Cancer genes registered in OncoKB database Table S3. Genes assessed by exome sequencing panel and recognized as cancer predisposition genes Table S4. Oligonucleotide sequences Table S5. Clinical features of the study cohort Table S6. Genetic profile of the study cohort Table S7. Clinical significance of SNV/Indels classified as Tier 1 or Tier 2 variants Table S8. Main features of Tier 1 potential germline variants Table S9. Clinical significance of CNVs, CNN‐LOH, and gene fusions classified as Tier 1 or Tier 2 Table S10. Tier 3 variants predicted as deleterious and reclassified as potentially pathogenic Table S11. Frequencies of genetic subtypes detected in this study and comparison with other cohorts Data Availability Statement The main data supporting the results of this study are available in the paper and its supplementary information. All data generated and datasets analyzed in the current study are available from the corresponding author upon reasonable request.