Clinical profiling of AML1::ETO and KIT exon 17 mutation in pediatric AML by high-throughput drug sensitivity Authors: Zhilin Hu, Xue Tang, Fen Chen, Tonghui Li, Yi Liu, Guichi Zhou, Qian Li, Shilin Liu, Ying Wang, Feiqiu Wen, Huirong Mai, Lulu Wang, Sixi Liu Published: 2026 ABSTRACT BackgroundThe favorable prognosis of pediatric AML1::ETO acute myeloid leukemia (AML) is well-established, yet the impact of co-occurring KIT mutations—particularly in exon 17—on clinical outcomes and chemosensitivity remains incompletely defined.MethodsWe analyzed the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) database comprising 957 pediatric AML patients to assess the clinical characteristics and prognosis of pediatric patients harboring AML1::ETO and KIT mutation using Kaplan–Meier survival analysis and Cox proportional hazards models. Subsequently, we performed experimental validation using bone marrow samples from 16 pediatric AML patients at Shenzhen Children's Hospital by high-throughput drug sensitivity (HDS) screening against therapeutic agents and whole transcriptome sequencing analysis of significant genes. Functional enrichment analysis of these differential expression genes was carried out by Gene Ontology.ResultsAnalysis from the TARGET database revealed that while AML1::ETO AML generally carries a favorable prognosis, concomitant KIT exon 17 mutations significantly attenuate this survival advantage. Exploratory experimental validation in a limited cohort of AML1::ETO and KIT exon 17 mutation patients suggested a potential trend of broad drug resistance such as cytarabine and daunorubicin. Preliminary transcriptomic analysis identified upregulation of SOCS1, a negative regulator of the JAK-STAT pathway, as a potential feature of this resistant phenotype. Furthermore, elevated SOCS1 expression was associated with poor prognosis.ConclusionsWe conclude that KIT mutations, especially exon 17, confer a high-risk phenotype in otherwise favorable pediatric AML1::ETO AML. Our exploratory data suggest this may be associated with a chemoresistant profile, potentially driven by SOCS1-associated JAK-STAT dysregulation. These findings highlight the necessity of refined risk stratification based on KIT exon profiling and support targeting the SOCS1/JAK-STAT axis to overcome therapy resistance.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12885-026-16289-4. Introduction As the leading pediatric malignancy, leukemia accounts for the highest proportion of cancer-related morbidity and mortality among children aged 0 −14 years in both China and globally [1, 2]. Core-binding factor (CBF) is a heterodimeric transcriptional regulator comprising four subunits with differential DNA-binding capacities, which are essential during embryonic hematopoietic development [3]. CBF-AML represents approximately 30% of pediatric AML cases and is designated as a favorable-risk subtype under both the World Health Organization (WHO) 2022 and International Consensus Classification (ICC) characterized by high initial complete remission rates [4–7]. CBF-AML is defined by two characteristic genetic alterations: the t(8;21)(q22;q22) translocation generating AML1::ETO fusion gene, and either inv(16)(p13.1q22) or t(16;16)(p13.1;q22) leading to the CBFB::MYH11 rearrangement. Although the prognosis is generally favorable, mutations in the KIT gene, particularly involving exons 8 and 17, occur in 10–42% of pediatric CBF-AML cases. Genomic studies indicate that exon 17 mutations (e.g., D816, N822, L793) lead to constitutive kinase activation, which correlates with worse outcomes, particularly in adult patients with t(8;21) AML [8–10]. However, the prognostic value of KIT mutations in pediatric t(8;21) AML is controversial. Previous studies suggested that pediatric CBF-AML harboring KIT mutations drives clonal persistence, minimal residual disease (MRD), and heightens the risk of relapse in pediatric cohorts [11–13]. Subsequent pediatric research has further associated these mutations with inferior overall survival (OS) and disease-free survival (DFS), specifically in t(8;21) patients, with negligible effects in the inv(16) subtype [14–16]. Conversely, several American and Chinese cohorts have reported no significant prognostic impact, suggesting that clinical outcomes may depend on specific treatment protocols, such as the inclusion of Gemtuzumab Ozogamicin (GO), which may abrogate the negative risk of KIT mutations [17, 18]. To address such clinical discrepancies, functional precision medicine approaches are increasingly valuable. High-throughput drug sensitivity (HDS) testing utilizes an enhanced conditional reprogramming method to rapidly grow primary cancer cells while maintaining their original drug response profiles and genetic stability. This renewable assay allows for the simultaneous screening of hundreds of drugs, aiding in personalized risk assessment and facilitating the translation of novel agents into precision therapies [19–21]. Notably, Wang et al. systematically integrated ex vivo drug sensitivity profiling with genomic analyses in 52 pediatric AML cases, identifying pediatric specific therapeutic vulnerabilities. Their study highlighted critical pharmacogenomic associations, such as KMT2C mutations serving as dual biomarkers for BCL-2 inhibitor sensitivity and adverse outcomes, providing foundational evidence for the utility of subtype-specific investigations in pediatric AML [20]. Despite these advances, the underlying biological mechanisms driving the prognosis of pediatric t(8;21) AML with KIT mutations, and their potential correlation with chemotherapy sensitivity, remain poorly understood. This study aims to elucidate the clinical characteristics of pediatric patients with t(8;21) AML harboring KIT mutations. Furthermore, we utilize HDS and whole transcriptome sequencing analysis to assess drug sensitivity, clarifying the underlying molecular mechanisms to advance targeted precision therapy. Materials and methods Patients and databases To comprehensively assess the impact of KIT mutations from clinical outcomes to biological mechanisms, we utilized two distinct patient cohorts. In the clinical cohort, retrospective clinical and genomic data were obtained from the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) AML initiative (https://ocg.cancer.gov/programs/target). This cohort aggregates de-identified multimodal data from patients enrolled in Children’s Oncology Group (COG) protocols AAML0531, AAML03P1, and CCG-2961. We included 957 pediatric patients diagnosed between 1996 and 2010, with follow-up data available through 2015. This dataset was employed to establish the prognostic value of concurrent AML1::ETO and KIT mutations through survival analysis. In the experimental cohort, we recruited a prospective cohort of 14 newly diagnosed pediatric AML patients treated at Shenzhen Children’s Hospital from 2018 to 2020 to elucidate the functional characterization and drug sensitivity profiling. Bone marrow (BM) samples were obtained from three patients with concurrent AML1::ETO and KIT exon 17 (D816V) mutations and 13 control patients lacking these alterations. Informed consent was obtained from all patients or their guardians. The study was approved by the Ethics Committee of Shenzhen Children’s Hospital (Approval No. 202111202) and conducted in accordance with the Declaration of Helsinki. Primary mononuclear cells were isolated from fresh BM samples via Ficoll density gradient centrifugation. High throughput drug sensitivity screening Pediatric AML patients with AML1::ETO and KIT exon 17 mutation (D816V site) and patients without AML1::ETO nor KIT exon 17 mutations were identified through transcriptome analysis. HDS testing before initiation of treatment was performed on the BM samples from patients by Hefei PreceDo Pharmaceuticals Co., Ltd. (Hefei, China) to evaluate their sensitivity to chemotherapy drugs and treatment regimens. To ensure a robust evaluation of both proliferative and quiescent leukemic populations, the HDS testing library included a balanced representation of classical cell-cycle-dependent chemotherapeutics and non-cycle-dependent targeted agents. Detailed information regarding the clinical dosage and tested concentration is provided in Table S1. Cells were exposed to compounds for 72 h in a humidified incubator at 37 °C with 5% CO₂. All drug concentrations and DMSO controls were tested in independent triplicates. Cell viability was assessed using the XTT assay. After 72 h, XTT labeling mixture with an electron-coupling reagent was added, followed by 4-h incubation at 37 °C in the dark. Absorbance was measured using a microplate reader and was proportional to the number of viable cells. Viability data were background-subtracted and normalized to DMSO-treated controls. The inhibition rate (%) was calculated as: [(OD_control − OD_drug)/(OD_control − OD_blank)] × 100%. Drug sensitivity was classified as follows: highly sensitive (≥ 70%), moderately sensitive (50–69%), low sensitivity (30–49%), and insensitive (< 30%). Whole Transcriptome Sequencing (WTS) Total RNA was extracted from samples using commercial RNA isolation KIT, and RNA integrity was assessed using a Bioanalyzer, with RIN ≥ 7.0 considered acceptable. Ribosomal RNA was depleted or poly(A)-selected before RNA fragmentation and cDNA synthesis. Libraries were prepared through end repair, A-tailing, adapter ligation, and PCR amplification, followed by size selection. Sequencing was performed on the Illumina platform, generating paired-end reads (100–150 bp). Raw reads were quality-checked with FastQC, trimmed with Cutadapt, and aligned to the reference genome using HISAT2/STAR. Transcript assembly and quantification were performed using StringTie/RSEM, followed by differential expression analysis with DESeq2/edgeR. Functional annotation and pathway enrichment analyses were conducted for biological interpretation. Differentially expressed genes screening and analysis The robust multi-array average (RMA) in R software was applied to explore the gene expression data. Differentially expressed genes (DEGs) between patients with AML1::ETO and KIT exon 17 mutations and patients without AML1::ETO nor KIT exon 17 mutations cases were identified using the Bayesian method by the linear models for microarray expression data (LIMMA) package in R software. |Log2 fold change (FC)|> 1 and a P value < 0.01 were considered as threshold points. All the above operations were run with scripts in the R software. The Ggplot2 package was conducted to show the heatmap. Construction of protein–protein interaction (PPI) network Protein–protein interaction (PPI) analysis of the up-regulated differentially expressed genes (up-DEGs) was performed using the STRING database (https://string-db.org/), applying a medium confidence interaction score of ≥ 0.4. The resulting network was visualized using Cytoscape (version 3.9.0). Key hub genes were identified via the CytoHubba plugin using the Density of Maximum Neighborhood Component (DMNC) and Maximal Clique Centrality (MCC) algorithms. The top 13 genes from each method were intersected to identify the core gene signature. Subsequently, Gene Ontology enrichment analysis was conducted using the DAVID database (http://david.ncifcrf.gov/), with a significance threshold of P < 0.05. To evaluate the prognostic significance of the hub genes, we utilized the Gene Expression Profiling Interactive Analysis (GEPIA) web server (http://gepia.cancer-pku.cn/) to analyze the LAML dataset of The Cancer Genome Atlas (TCGA). Patients were stratified into high- and low-expression groups using median transcript level of the target gene for survival analysis [22]. Statistical analysis Categorical variables were compared using the chi-square test and Fisher's exact test where appropriate to evaluate the distribution of baseline characteristics and associations between variables. Kaplan–Meier analysis and Log-rank test were employed to estimate survival probabilities and visualize the survival distribution of patients over time. To identify independent prognostic factors for survival, a Cox proportional hazards model was applied, with variables showing P ≤ 0.2 in univariate analysis selected for inclusion in the multivariate model. P < 0.05 was considered statistically significant in the multivariate model. The Cox model provided hazard ratios (HRs) and 95% confidence intervals (CIs) to quantify the impact of each prognostic factor on survival outcomes. Analyses of clinical characteristics and prognosis were conducted using SPSS, and additional data visualization, such as forest plots for HRs, was performed using Sanger Box [23]. Results Clinical and prognostic characteristics of AML1::ETO AML Favorable prognosis of AML1::ETO AML Our comparative analysis of clinical outcomes revealed that pediatric patients harboring the AML1::ETO fusion exhibited significantly better treatment responses and survival rates than those lacking this genetic alteration. Compared with patients without AML1::ETO, a significantly higher complete remission (CR) rate was observed in the AML1::ETO group after the second course of chemotherapy (94.0% vs. 84.1%, P < 0.01). More importantly, the AML1::ETO group demonstrated significantly lower levels of minimal residual disease (MRD) after both course 1 (21.9% vs. 34.3%, P = 0.02) and course 2 (9.9% vs. 20.6%, P = 0.02), which was shown in Table 1. This advantage in early treatment response translated into significantly improved long-term survival outcomes. As shown in Fig. 1A and B, patients with AML1::ETO AML had significantly better event-free survival (EFS) with HR = 0.48 (95% CI: 0.35–0.65, P < 0.001) and OS with HR = 0.48 (95% CI: 0.33–0.71, P < 0.001) than those without AML1::ETO.Table 1Therapeutic evaluation according to AML subtype in TARGETAML1::ETOWithout AML1::ETOP valueCourse 1 Bone marrow0.05CR81.3% (109/134)73.5% (596/811)Not CR18.7% (25/134)26.5% (215/811) MRD0.02*Yes21.9% (21/96)34.3% (222/648)No78.1% (75/96)65.7% (426/648)Course 2 Bone marrow< 0.01**CR94.0% (125/133)84.1% (668/794)Not CR6.0% (8/133)15.9% (126/794) MRD0.02*Yes9.9% (9/91)20.6% (113/549)No90.1% (82/91)79.4% (436/549)P values were analyzed by the Chi-Square Test, compared with AML1::ETO groupCR Complete remission, MRD Minimal residual disease*P < 0.05**P < 0.01Fig. 1Survival in patients with AML subtypes in TARGET, AML1::ETO versus Without AML1::ETO. ​Kaplan–Meier curves compare survival outcomes between genetic subtypes. A EFS and B​ OS of patients with AML1::ETO (light blue) versus without AML1::ETO AML (red). HR with 95% confidence intervals are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% confidence intervals for survival curves. Numbers at risk at sequential time points are shown below each panel. EFS, event-free survival; OS, overall survival; AML, acute myeloid leukemia; CBF, core-binding factor; HR, hazard ratios Table 1Therapeutic evaluation according to AML subtype in TARGETAML1::ETOWithout AML1::ETOP valueCourse 1 Bone marrow0.05CR81.3% (109/134)73.5% (596/811)Not CR18.7% (25/134)26.5% (215/811) MRD0.02*Yes21.9% (21/96)34.3% (222/648)No78.1% (75/96)65.7% (426/648)Course 2 Bone marrow< 0.01**CR94.0% (125/133)84.1% (668/794)Not CR6.0% (8/133)15.9% (126/794) MRD0.02*Yes9.9% (9/91)20.6% (113/549)No90.1% (82/91)79.4% (436/549)P values were analyzed by the Chi-Square Test, compared with AML1::ETO groupCR Complete remission, MRD Minimal residual disease*P < 0.05**P < 0.01 Therapeutic evaluation according to AML subtype in TARGET P values were analyzed by the Chi-Square Test, compared with AML1::ETO group CR Complete remission, MRD Minimal residual disease *P < 0.05 **P < 0.01 Survival in patients with AML subtypes in TARGET, AML1::ETO versus Without AML1::ETO. ​Kaplan–Meier curves compare survival outcomes between genetic subtypes. A EFS and B​ OS of patients with AML1::ETO (light blue) versus without AML1::ETO AML (red). HR with 95% confidence intervals are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% confidence intervals for survival curves. Numbers at risk at sequential time points are shown below each panel. EFS, event-free survival; OS, overall survival; AML, acute myeloid leukemia; CBF, core-binding factor; HR, hazard ratios KIT mutation attenuate the favorable prognostic impact of AML1::ETO AML​​ Despite the overall favorable outlook for AML1::ETO AML, the presence of KIT mutations appears to compromise this advantage. When analyzing the subset of AML patients with KIT mutations, those harboring concurrent AML1::ETO and KIT mutations exhibited a higher MRD positivity rate (30.0%) after the first induction course compared to patients without AML1::ETO but with KIT mutations (4.0%, P = 0.05), which was shown in Table 2. As shown in Fig. S1, the co-occurrence of KIT mutations with AML1::ETO resulted in the EFS and OS were not distinguishable from those of patients lacking AML1::ETO but with KIT mutations, suggesting that the presence of KIT mutations may mitigate the otherwise favorable prognostic effect of AML1::ETO.Table 2Therapeutic evaluation according to AML subtype with KIT mutation in TARGETWith AML1::ETO and KIT mutationWithout AML1::ETO but with KIT mutationP valueCourse 1 Bone marrow0.35 CR78.6% (22/28)90.6% (29/32) Not CR21.4% (6/28)9.4% (3/32) MRD0.05 Yes30.0% (6/20)4.0% (1/25) No70.0% (14/20)96.0% (24/25)Course 2 Bone marrow0.48 CR96.4% (27/28)100.0% (31/31) Not CR3.6% (1/28)0.0% (0/31) MRD0.44 Yes5.0% (1/20)0.0% (0/25) No95.0% (19/20)100.0% (25/25)P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual disease Table 2Therapeutic evaluation according to AML subtype with KIT mutation in TARGETWith AML1::ETO and KIT mutationWithout AML1::ETO but with KIT mutationP valueCourse 1 Bone marrow0.35 CR78.6% (22/28)90.6% (29/32) Not CR21.4% (6/28)9.4% (3/32) MRD0.05 Yes30.0% (6/20)4.0% (1/25) No70.0% (14/20)96.0% (24/25)Course 2 Bone marrow0.48 CR96.4% (27/28)100.0% (31/31) Not CR3.6% (1/28)0.0% (0/31) MRD0.44 Yes5.0% (1/20)0.0% (0/25) No95.0% (19/20)100.0% (25/25)P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual disease Therapeutic evaluation according to AML subtype with KIT mutation in TARGET P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation group CR Complete remission, MRD Minimal residual disease Furthermore, we evaluate the prognostic significance of KIT mutations in patients with AML1::ETO. Therapeutic responses of AML1::ETO patients stratified by KIT mutation in Table 3. In AML1::ETO, patients harboring KIT mutations exhibited a numerically higher MRD positivity rate compared to those without KIT mutations (30.0% vs. 19.4%), though the difference did not reach statistical significance (P = 0.48). Survival analysis was performed to evaluate the impact of KIT mutation sites on EFS and OS in AML1::ETO. Shown in Fig. 2A and B, overall presence of KIT mutations did not significantly affect EFS or OS in AML1::ETO patients. Given the functional heterogeneity of different mutation sites, we further stratified the cohort by specific exons. Notably, among the various mutation sites including exon 8 and exon 17, only patients harboring KIT exon 17 mutations exhibited a potential trend toward inferior survival outcomes when compared to those without KIT mutations (Fig. 2C and D). Although these survival differences did not reach formal statistical significance, the elevated hazard ratios (HR = 2.09, 95% CI: 0.95–4.57, P = 0.06 for EFS; HR = 2.25, 95% CI: 0.90–5.62, P = 0.07 for OS) suggest a clinically meaningful adverse impact. As shown in Fig. 2E and G, the univariate analysis and multivariate analysis of EFS in AML1::ETO patients were conducted, no statistically significant predictor was noticed. Crucially, parallel analyses were conducted for OS in Fig. 2F and H, the multivariable Cox regression model conclusively demonstrated that the presence of a KIT exon 17 mutation acts as a potential independent adverse prognostic factor for OS (HR = 3.61, 95% CI: 1.25–10.48, P = 0.02).Table 3Therapeutic evaluation according to KIT mutation with AML1::ETO in TARGETAML1::ETO with KIT mutationAML1::ETO without KIT mutationP valueCourse 1 Bone marrow0.81 CR78.6% (22/28)80.6% (79/98) Not CR21.4% (6/28)19.4% (19/98) MRD0.48 Yes30.0% (6/20)19.4% (14/72) No70.0% (14/20)80.6% (58/72)Course 2 Bone marrow0.95 CR96.4% (27/28)93.8% (91/97) Not CR3.6% (1/28)6.2% (6/97) MRD0.77 Yes5.0% (1/20)10.4% (7/67) No95.0% (19/20)89.6% (60/67)P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual diseaseFig. 2Survival analysis of AML1::ETO patients based on KIT mutation in TARGET. Kaplan–Meier curves compare survival outcomes of the patients with AML1::ETO based on KIT mutation in TARGET. A EFS and B​ OS of AML1::ETO patients with KIT mutation (light blue) versus without KIT mutation (red). C EFS and D OS of AML1::ETO patients with Exon 8 mutation (light blue), Exon 17 mutation (light green), Exon 8 + 17 mutation (dark blue), versus without KIT mutation (red). Cox proportional hazards regression was performed to evaluate the impact of clinical and treatment variables in AML1::ETO. HR with 95% CIs are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% CIs for survival curves. Numbers at risk at sequential time points are shown below each panel. Univariate Cox regression analysis of (E)​ EFS and (F) OS showing HR and 95% confidence intervals for gender, race, ethnicity, age, treatment protocol, initial WBC count, FAB classification, treatment response after course 1, GO treatment, and relapse/induction failure in AML1::ETO patients. ​ Multivariate Cox regression analysis of (G) EFS and (H) OS incorporating the variables with P ≤ 0.2. HR > 1 indicates increased risk, HR < 1 indicates protective effect. Statistical significance was determined using Wald tests. EFS, event-free survival; OS, overall survival; HR, hazard ratios; WBC, white blood cell; FAB, French-American-British; MRD, minimal residual disease; GO, gemtuzumab ozogamicin; HR, hazard ratios; CI, confidence interval Table 3Therapeutic evaluation according to KIT mutation with AML1::ETO in TARGETAML1::ETO with KIT mutationAML1::ETO without KIT mutationP valueCourse 1 Bone marrow0.81 CR78.6% (22/28)80.6% (79/98) Not CR21.4% (6/28)19.4% (19/98) MRD0.48 Yes30.0% (6/20)19.4% (14/72) No70.0% (14/20)80.6% (58/72)Course 2 Bone marrow0.95 CR96.4% (27/28)93.8% (91/97) Not CR3.6% (1/28)6.2% (6/97) MRD0.77 Yes5.0% (1/20)10.4% (7/67) No95.0% (19/20)89.6% (60/67)P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual disease Therapeutic evaluation according to KIT mutation with AML1::ETO in TARGET P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation group CR Complete remission, MRD Minimal residual disease Survival analysis of AML1::ETO patients based on KIT mutation in TARGET. Kaplan–Meier curves compare survival outcomes of the patients with AML1::ETO based on KIT mutation in TARGET. A EFS and B​ OS of AML1::ETO patients with KIT mutation (light blue) versus without KIT mutation (red). C EFS and D OS of AML1::ETO patients with Exon 8 mutation (light blue), Exon 17 mutation (light green), Exon 8 + 17 mutation (dark blue), versus without KIT mutation (red). Cox proportional hazards regression was performed to evaluate the impact of clinical and treatment variables in AML1::ETO. HR with 95% CIs are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% CIs for survival curves. Numbers at risk at sequential time points are shown below each panel. Univariate Cox regression analysis of (E)​ EFS and (F) OS showing HR and 95% confidence intervals for gender, race, ethnicity, age, treatment protocol, initial WBC count, FAB classification, treatment response after course 1, GO treatment, and relapse/induction failure in AML1::ETO patients. ​ Multivariate Cox regression analysis of (G) EFS and (H) OS incorporating the variables with P ≤ 0.2. HR > 1 indicates increased risk, HR < 1 indicates protective effect. Statistical significance was determined using Wald tests. EFS, event-free survival; OS, overall survival; HR, hazard ratios; WBC, white blood cell; FAB, French-American-British; MRD, minimal residual disease; GO, gemtuzumab ozogamicin; HR, hazard ratios; CI, confidence interval To better understand the unique biological vulnerabilities driving this distinct clinical resistance, we subsequently focused on the group with concurrent AML1::ETO and KIT exon 17 mutations. Clinical characteristics and prognostic indicators of AML1::ETO and KIT exon 17 mutations Clinical features and outcomes of pediatric patients with AML1::ETO and KIT exon 17 mutation (n = 15) against a control group of patients without AML1::ETO nor KIT mutation (n = 81), with details shown in Table 4.Table 4Clinical characteristics of patients with AML1::ETO and KIT exon 17 mutation and without AML1::ETO nor KIT mutation in TARGETTotalAML1::ETO and KIT exon 17 mutationWithout AML1::ETO nor 17 KIT mutationP valuePatients, n(%)961581Demographics Gender0.99Male46.9% (45/96)46.7% (7/15)46.9% (38/81)Female53.1% (51/96)53.3% (8/15)53.1% (43/81) Race0.001**Asian2.1% (2/96)6.7% (1/15)1.2% (1/81)Black or African8.3% (8/96)0.0% (0/15)9.9% (8/81)White72.9% (70/96)46.7% (7/15)77.8% (63/81)Other9.4% (9/96)13.3% (2/15)8.7% (7/81)Unknown6.3% (6/96)33.3% (5/15)1.2% (1/81)American Indian or Alaska Native1.0% (1/96)0.0% (0/15)1.2% (1/81) Ethnicity0.01*Hispanic or Latino26.1% (25/96)53.3% (8/15)21.0% (17/81)Not Hispanic or Latino70.8% (68/96)40.0% (6/15)76.5% (62/81)Unknown3.1% (3/96)6.7% (1/15)2.5% (2/81) Age(years)0.78< 1039.6% (38/96)33.3% (5/15)40.7% (33/81)≥ 1060.4% (58/96)66.7% (10/15)59.3% (48/81) Risk group1.00High2.1% (2/96)0.0% (0/15)2.5% (2/81)Standard2.1% (2/96)0.0% (0/15)2.5% (2/81)Low95.8% (92/96)100.0% (15/15)95.0% (77/81)Clinical features Treatment protocol0.66AAML03P17.3% (7/96)0.0% (0/15)8.6% (7/81)AAML053187.5% (84/96)93.3% (14/15)86.4% (70/81)CCG-29615.2% (5/96)6.7% (1/15)5.0% (4/81) Initial WBC count(10^9/L)0.004**< 5039.6% (38/96)73.3% (11/15)33.3% (27/81)≥ 5060.4% (58/96)26.7% (4/15)66.7% (54/81) FAB Category< 0.001***M13.1% (3/96)13.3% (2/15)1.2% (1/81)M212.5% (12/96)73.4% (11/15)1.2% (1/81)M473.0 % (70/96)0.0% (0/15)86.5% (70/81)M51.0% (1/96)0.0% (0/15)1.2% (1/81)NOS1.0% (1/96)0.0% (0/15)1.2% (1/81)Unknown9.4% (9/96)13.3% (2/15)8.7% (7/81)Therapeutic evaluation Course 1Bone marrow0.38CR84.2% (80/95)73.3% (11/15)86.2% (69/80)Not CR15.8% (15/95)26.7% (4/15)13.8% (11/80)MRD0.52Yes13.6% (11/81)23.1% (3/13)11.8% (8/68)No86.4% (70/81)76.9% (10/13)88.2% (60/68) Course 2Bone marrow1.00CR93.6% (88/94)93.3% (14/15)93.7% (74/79)Not CR6.4% (6/94)6.7% (1/15)6.3% (5/79)MRD1.00Yes6.0% (4/67)9.1% (1/11)5.4% (3/56)No94.0% (63/67)90.9% (10/11)94.6% (53/56)GO treatment0.34Yes54.9% (50/91)42.9% (6/14)57.1% (44/77)No45.1% (41/91)57.1% (8/14)42.9% (33/77)Relapse/Induction Failure Bone Marrow Site of Relapse/Induction Failure85.7% (30/35)85.7% (6/7)85.7% (24/28)1.00 CNS Site of Relapse/Induction Failure34.3% (12/35)28.6% (2/7)35.7% (10/28)1.00 Chloroma Site of Relapse/Induction Failure5.7% (2/35)14.3% (1/7)3.6% (1/28)0.86 Cytogenetic Site of Relapse/Induction Failure14.3% (5/35)0.0% (0/7)17.9% (5/28)0.56 Other Site of Relapse/Induction Failure5.7% (2/35)0.0% (0/7)7.1% (2/28)1.00P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual disease, WBC White blood cell, GO gemtuzumab ozogamicin* P < 0.05** P < 0.01*** P < 0.001 Table 4Clinical characteristics of patients with AML1::ETO and KIT exon 17 mutation and without AML1::ETO nor KIT mutation in TARGETTotalAML1::ETO and KIT exon 17 mutationWithout AML1::ETO nor 17 KIT mutationP valuePatients, n(%)961581Demographics Gender0.99Male46.9% (45/96)46.7% (7/15)46.9% (38/81)Female53.1% (51/96)53.3% (8/15)53.1% (43/81) Race0.001**Asian2.1% (2/96)6.7% (1/15)1.2% (1/81)Black or African8.3% (8/96)0.0% (0/15)9.9% (8/81)White72.9% (70/96)46.7% (7/15)77.8% (63/81)Other9.4% (9/96)13.3% (2/15)8.7% (7/81)Unknown6.3% (6/96)33.3% (5/15)1.2% (1/81)American Indian or Alaska Native1.0% (1/96)0.0% (0/15)1.2% (1/81) Ethnicity0.01*Hispanic or Latino26.1% (25/96)53.3% (8/15)21.0% (17/81)Not Hispanic or Latino70.8% (68/96)40.0% (6/15)76.5% (62/81)Unknown3.1% (3/96)6.7% (1/15)2.5% (2/81) Age(years)0.78< 1039.6% (38/96)33.3% (5/15)40.7% (33/81)≥ 1060.4% (58/96)66.7% (10/15)59.3% (48/81) Risk group1.00High2.1% (2/96)0.0% (0/15)2.5% (2/81)Standard2.1% (2/96)0.0% (0/15)2.5% (2/81)Low95.8% (92/96)100.0% (15/15)95.0% (77/81)Clinical features Treatment protocol0.66AAML03P17.3% (7/96)0.0% (0/15)8.6% (7/81)AAML053187.5% (84/96)93.3% (14/15)86.4% (70/81)CCG-29615.2% (5/96)6.7% (1/15)5.0% (4/81) Initial WBC count(10^9/L)0.004**< 5039.6% (38/96)73.3% (11/15)33.3% (27/81)≥ 5060.4% (58/96)26.7% (4/15)66.7% (54/81) FAB Category< 0.001***M13.1% (3/96)13.3% (2/15)1.2% (1/81)M212.5% (12/96)73.4% (11/15)1.2% (1/81)M473.0 % (70/96)0.0% (0/15)86.5% (70/81)M51.0% (1/96)0.0% (0/15)1.2% (1/81)NOS1.0% (1/96)0.0% (0/15)1.2% (1/81)Unknown9.4% (9/96)13.3% (2/15)8.7% (7/81)Therapeutic evaluation Course 1Bone marrow0.38CR84.2% (80/95)73.3% (11/15)86.2% (69/80)Not CR15.8% (15/95)26.7% (4/15)13.8% (11/80)MRD0.52Yes13.6% (11/81)23.1% (3/13)11.8% (8/68)No86.4% (70/81)76.9% (10/13)88.2% (60/68) Course 2Bone marrow1.00CR93.6% (88/94)93.3% (14/15)93.7% (74/79)Not CR6.4% (6/94)6.7% (1/15)6.3% (5/79)MRD1.00Yes6.0% (4/67)9.1% (1/11)5.4% (3/56)No94.0% (63/67)90.9% (10/11)94.6% (53/56)GO treatment0.34Yes54.9% (50/91)42.9% (6/14)57.1% (44/77)No45.1% (41/91)57.1% (8/14)42.9% (33/77)Relapse/Induction Failure Bone Marrow Site of Relapse/Induction Failure85.7% (30/35)85.7% (6/7)85.7% (24/28)1.00 CNS Site of Relapse/Induction Failure34.3% (12/35)28.6% (2/7)35.7% (10/28)1.00 Chloroma Site of Relapse/Induction Failure5.7% (2/35)14.3% (1/7)3.6% (1/28)0.86 Cytogenetic Site of Relapse/Induction Failure14.3% (5/35)0.0% (0/7)17.9% (5/28)0.56 Other Site of Relapse/Induction Failure5.7% (2/35)0.0% (0/7)7.1% (2/28)1.00P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation groupCR Complete remission, MRD Minimal residual disease, WBC White blood cell, GO gemtuzumab ozogamicin* P < 0.05** P < 0.01*** P < 0.001 Clinical characteristics of patients with AML1::ETO and KIT exon 17 mutation and without AML1::ETO nor KIT mutation in TARGET P values were analyzed by the Chi-Square Test, compared with AML1::ETO and KIT mutation group CR Complete remission, MRD Minimal residual disease, WBC White blood cell, GO gemtuzumab ozogamicin * P < 0.05 ** P < 0.01 *** P < 0.001 Notable differences were observed in the presenting clinical features. Patients with AML1::ETO and KIT exon 17 mutation had a significantly lower initial white blood cell (WBC) count compared to the control group (P = 0.004), with a majority (73.3%) presenting with WBC < 50 × 10⁹/L. Furthermore, the AML1::ETO and KIT mutation group was strongly associated with the M2 morphology according to the French-American-British (FAB) classification system (73.4%), whereas without AML1::ETO nor KIT mutation group was predominantly classified as FAB M4 (86.5%, P < 0.001). Critically, despite these distinct baseline profiles, the presence of KIT mutation in AML1::ETO AML appeared to attenuate its typically favorable prognosis. As shown in Table 4, there were no statistically significant differences in MRD response and CR rate between the two groups in 2 courses of chemotherapy. However, following induction therapy in course 1, patients harboring concurrent AML1::ETO and KIT exon 17 mutations exhibited a notably higher prevalence of persistent MRD of 23.1% accompanied with lower CR rate of 73.3%. Interestingly, this early suboptimal response was largely overcome with further treatment, as both groups achieved highly comparable and excellent CR rates by the end of Course 2. Furthermore, survival analysis was conducted to explore long term survival outcome in Fig. 3A and B. While the difference in EFS was not statistically significant, the AML1::ETO and KIT exon 17 mutation group demonstrated a trend toward inferior OS compared to the without AML1::ETO nor KIT exon 17 mutation group (Fig. 3B). Although no significance was noticed (P = 0.07), it indicates a potential adverse impact for OS (HR = 2.41, 95% CI: 0.93–6.22).Fig. 3Survival analysis of AML1::ETO with KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation in TARGET. Kaplan–Meier curves compare survival outcomes of the patients with AML1::ETO with KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation in TARGET. A​ EFS and B​ OS of AML1::ETO patients with KIT exon 17 mutation (light blue) versus without AML1::ETO nor KIT exon 17 mutation (red). Cox proportional hazards regression was performed to evaluate the impact of clinical and treatment variables in AML1::ETO. HR with 95% CIs are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% CIs for survival curves. Numbers at risk at sequential time points are shown below each panel. Univariate Cox regression analysis of (C)​ EFS and (D) OS showing HR and 95% confidence intervals for gender, race, ethnicity, age, treatment protocol, initial WBC count, FAB classification, treatment response after course 1, GO treatment, and relapse/induction failure in patients with AML1::ETO with KIT exon 17 mutation. ​ Multivariate Cox regression analysis of (E) EFS and (F) OS incorporating the variables with P ≤ 0.2. HR > 1 indicates increased risk, HR < 1 indicates protective effect. Statistical significance was determined using Wald tests. EFS, event-free survival; OS, overall survival; HR, hazard ratios; WBC, white blood cell; FAB, French-American-British; MRD, minimal residual disease; GO, gemtuzumab ozogamicin; HR, hazard ratios; CI, confidence interval Survival analysis of AML1::ETO with KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation in TARGET. Kaplan–Meier curves compare survival outcomes of the patients with AML1::ETO with KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation in TARGET. A​ EFS and B​ OS of AML1::ETO patients with KIT exon 17 mutation (light blue) versus without AML1::ETO nor KIT exon 17 mutation (red). Cox proportional hazards regression was performed to evaluate the impact of clinical and treatment variables in AML1::ETO. HR with 95% CIs are indicated; log-rank tests were used for statistical comparison. Gray shaded areas represent 95% CIs for survival curves. Numbers at risk at sequential time points are shown below each panel. Univariate Cox regression analysis of (C)​ EFS and (D) OS showing HR and 95% confidence intervals for gender, race, ethnicity, age, treatment protocol, initial WBC count, FAB classification, treatment response after course 1, GO treatment, and relapse/induction failure in patients with AML1::ETO with KIT exon 17 mutation. ​ Multivariate Cox regression analysis of (E) EFS and (F) OS incorporating the variables with P ≤ 0.2. HR > 1 indicates increased risk, HR < 1 indicates protective effect. Statistical significance was determined using Wald tests. EFS, event-free survival; OS, overall survival; HR, hazard ratios; WBC, white blood cell; FAB, French-American-British; MRD, minimal residual disease; GO, gemtuzumab ozogamicin; HR, hazard ratios; CI, confidence interval Cox proportional hazards models were utilized to evaluate potential prognostic factors of patients with AML1::ETO and KIT exon 17 mutations, with variables achieving P ≤ 0.2 in the univariate analysis selected for the multivariate model. In the univariate analysis for EFS (Fig. 3C), initial WBC count and GO treatment met the inclusion criteria. Subsequent multivariate analysis (Fig. 3E) revealed that neither GO treatment nor initial WBC count is an isolated prognostic factor of EFS. For OS, age and treatment protocol were identified as potential predictors in the univariate analysis (Fig. 3D). Strikingly, in the final multivariate Cox model (Fig. 3F), both variables emerged as statistically significant independent prognostic factors. Older age (≥ 10 years) demonstrated a robust protective effect for OS (HR = 0.08, 95% CI: 0.01–0.74, P = 0.03). Additionally, treatment with the AAML0531 protocol was independently associated with significantly superior OS outcomes compared to the CCG-2961 protocol (HR = 0.04, 95% CI: 0.00–0.75, P = 0.03). High-throughput drug sensitivity test for pediatric leukemia patients To elucidate the chemotherapeutic response profiles of pediatric AML patients with AML1::ETO and KIT exon 17 mutations, we performed HDS testing on primary cells from newly diagnosed patients at our center, quantitatively determining their proliferation inhibition rates. The results showed that, compared with the control group, patients with AML1::ETO and KIT exon 17 mutation (D816V site) were resistant to multiple combination chemotherapy regimens for AML (Fig. 4A). Monotherapy analysis revealed that these patients were resistant to classic myeloid chemotherapy agents, including cytarabine, fludarabine, homoharringtonine, daunorubicin, mitoxantrone, and cladribine (Fig. 4B), suggesting that patients with this genetic profile may have poor responses to chemotherapy.Fig. 4Profiling of drug sensitivity in primary leukemic cells. A High-throughput screening results comparing the sensitivity of the patients of AML1::ETO and KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation group to a panel of combination chemotherapy regimens. B High throughput drug sensitivity of Fludarabine, Cytarabine, Homoharringtonine, Daunorubicin, Mitoxantrone, and Cladribine of bone marrow in patients with AML1::ETO and KIT exon 17 mutation compared to the control. Statistical significance was indicated as follows: * P < 0.05, ** P < 0.01, *** P < 0.001 Profiling of drug sensitivity in primary leukemic cells. A High-throughput screening results comparing the sensitivity of the patients of AML1::ETO and KIT exon 17 mutation and without AML1::ETO nor KIT exon 17 mutation group to a panel of combination chemotherapy regimens. B High throughput drug sensitivity of Fludarabine, Cytarabine, Homoharringtonine, Daunorubicin, Mitoxantrone, and Cladribine of bone marrow in patients with AML1::ETO and KIT exon 17 mutation compared to the control. Statistical significance was indicated as follows: * P < 0.05, ** P < 0.01, *** P < 0.001 RNA-seq analysis for the AML1::ETO and KIT mutation vs. without AML1::ETO nor KIT mutation patients We further compared the gene expression of patients with AML1::ETO and KIT exon 17 mutation to those without such rearrangement and mutation. The most upregulated DEGs were marked in the volcano map, such as RUNX1T1 and HIF3A (Fig. 5A). To gain insight into the valuable hub genes in the upregulated DEGs, the protein–protein interaction network were examined. The top 50 hub genes were identified using both the DMNC and MCC algorithms. The intersection of the genes obtained from both algorithms includes the following 13 genes: NFKBIZ, JUNB, TNFAIP3, NFKB2, TEK, SOCS1, EIF2AK3, ZFP36, CXCL2, BCL3, IRF1, SOCS3, and FGF6 (Fig. 5B). Gene ontology analysis showed SOCS1, SOCS3 which in the up-regulated gene were highly associated with many important pathways, including negative regulation of receptor signaling pathway via JAK-STAT and cytokine-mediated signaling pathway (Fig. 5C). Further Kaplan–Meier analysis of hub genes demonstrated that SOCS1 was the only gene significantly associated with poor prognosis in AML (HR = 2.2, P = 0.0047; Fig. 5D). In contrast, neither SOCS3 alone (P = 0.68) nor a combined SOCS1/SOCS3 two-gene signature (P = 0.91) showed prognostic significance (Fig. S2). These findings support SOCS1 as an independent prognostic biomarker and suggest that SOCS1-associated dysregulation of the JAK–STAT pathway may contribute to AML pathogenesis.Fig. 5Bioinformatic analysis identifies SOCS1 as a key prognostic gene. A Volcano plot displaying differentially expressed genes (DEGs). A total of 1909 significant DEGs were identified between the Mutated and Control groups, comprising 1130 upregulated (red) and 779 downregulated (green) genes. B Protein–protein interaction network analysis. Hub genes were identified by taking the intersection of results from the DMNC and MCC algorithm. C Gene Ontology enrichment pathway analyses of the upregulated DEGs. D Kaplan–Meier survival analysis of SOCS1 expression in the TCGA-LAML cohort Bioinformatic analysis identifies SOCS1 as a key prognostic gene. A Volcano plot displaying differentially expressed genes (DEGs). A total of 1909 significant DEGs were identified between the Mutated and Control groups, comprising 1130 upregulated (red) and 779 downregulated (green) genes. B Protein–protein interaction network analysis. Hub genes were identified by taking the intersection of results from the DMNC and MCC algorithm. C Gene Ontology enrichment pathway analyses of the upregulated DEGs. D Kaplan–Meier survival analysis of SOCS1 expression in the TCGA-LAML cohort Discussion Our study integrates large-scale clinical profiling with functional drug sensitivity assays to unravel the paradox of pediatric AML1::ETO AML harboring KIT mutations. While AML1::ETO leukemia is traditionally classified as a favorable-risk entity, we confirmed that this survival advantage is significantly attenuated by the co-occurrence of KIT mutations, particularly those affecting exon 17. Crucially, our exploratory HDS profiling suggested a potential trend of chemoresistant phenotype in this subset, characterized by reduced sensitivity to standard induction agents in vitro. Transcriptomic analysis identified the upregulation of SOCS1 as a potential molecular feature of this resistant phenotype, suggesting that immune evasion via JAK-STAT dysregulation may drive the poor outcomes observed in these patients. Our study integrates large-scale clinical profiling with functional drug sensitivity assays to address the paradox of pediatric AML1::ETO AML harboring KIT exon 17 mutations. While AML1::ETO leukemia is traditionally classified as a favorable-risk entity, we demonstrate that this survival advantage is markedly attenuated by the co-occurrence of KIT mutations, especially in exon 17. Notably, our exploratory HDS profiling suggests a trend toward a chemoresistant phenotype in this subgroup, characterized by reduced in vitro sensitivity to standard induction agents. Transcriptomic analysis further identified upregulation of SOCS1 as a potential molecular feature of this resistant phenotype, implicating dysregulated JAK–STAT signaling and immune evasion in the poor outcomes observed. Consistent with the 2022 European Leukemia Net (ELN) classification, analysis of the TARGET cohort reaffirmed the high intrinsic chemosensitivity and favorable baseline prognosis of pediatric AML1::ETO AML [4, 24]. However, this favorable prognosis was significantly compromised in the presence of KIT mutations, in agreement with previous studies demonstrating that KIT mutations act as adverse prognostic markers in AML1::ETO AML, increasing relapse risk and reducing survival in pediatric patients, while exerting minimal impact in inv(16) AML subtypes [15, 25–27]. Qiu et al. further showed that KIT mutations independently predict higher cumulative incidence of relapse (CIR) and inferior OS in pediatric AML1::ETO AML [16]. Importantly, despite inferior survival, most AML1::ETO patients with KIT exon 17 mutations in TARGET cohort achieved complete remission after induction therapy, indicating that this subgroup is not primarily refractory. Instead, the adverse prognosis appears to stem from impaired early leukemic clearance kinetics. The observed suboptimal in vitro chemosensitivity suggests an intrinsic biological tolerance driven by KIT mutations. Although this tolerance does not preclude eventual morphological remission, it significantly delays early clonal eradication, resulting in MRD persistence. Delayed blast clearance during induction is a well-established indicator of intrinsic disease tolerance, and persistent MRD serves as a critical determinant of relapse risk in otherwise favorable t(8;21) AML [28–31]. Furthermore, our data indicate that this prognostic detriment is highly mutation-site specific. Significant survival differences were observed in patients harboring KIT exon 17 mutations, whereas exon 8 mutations showed no statistically significant impact. This is consistent with prior studies demonstrating that KIT mutations—particularly exon 17 alterations or minor subclones—are associated with delayed MRD clearance [13, 32]. These findings are further supported by the JCCG AML-12 trial, which identified KIT exon 17 mutations as an independent determinant of poor prognosis in pediatric AML1::ETO AML. Similarly, in adult cohorts, exon 17 mutations were associated with inferior relapse-free survival (RFS) and OS, whereas exon 8 mutations were not [33]. The prognostic disparity between exon 17 and exon 8 mutations likely reflects their distinct functional consequences: exon 17 mutations (e.g., D816V) induce constitutive, ligand-independent kinase activation, acting as potent oncogenic drivers, whereas exon 8 mutations primarily confer ligand hypersensitivity [34–36]. These findings highlight the necessity of exon-level molecular stratification to refine risk assessment. Our transcriptomic analysis identified RUNX1T1 and HIF3A among the most significantly upregulated DEGs. The prominent upregulation of RUNX1T1 provides robust internal validation for our cohort, representing the core component of the defining AML1::ETO fusion. More importantly, the marked enrichment of HIF3A highlights a critical role for hypoxia-adaptation mechanisms in this molecular subgroup. The hypoxic bone marrow microenvironment, primarily orchestrated by hypoxia-inducible factor (HIF) signaling networks, is increasingly recognized to protect leukemic stem cells from cytarabine-based chemotherapy, thereby promoting niche-mediated therapeutic tolerance [37–39]. Recent functional studies have demonstrated that inhibiting the HIF pathway directly sensitizes AML clones to standard induction regimens [40]. Further research is needed to verify the connection between HIF3A-associated hypoxic adaptation, KIT mutations, and the delayed early leukemic clearance observed in this cohort. We further identified SOCS1 upregulation as a potential driver of poor survival. As a key negative regulator of JAK–STAT signaling, SOCS1 may contribute to treatment resistance through mechanisms distinct from classical metabolic chemoresistance. Rather than sensitizing leukemic cells to JAK inhibition, elevated SOCS1 expression likely promotes immune evasion by suppressing T cell–mediated cytotoxicity [41, 42]. Emerging evidence further supports SOCS1 as an immune checkpoint regulator that enables leukemic cells to evade immune surveillance [43–45]. Notably, conventional ex vivo drug assays lack a functional immune microenvironment and therefore fail to capture these immune-mediated survival dynamics [46]. These findings suggest that SOCS1-driven poor outcomes may primarily reflect immune escape rather than intrinsic drug resistance. Future translational efforts must move beyond in vitro JAK inhibition and employ immune-competent in vivo models to evaluate whether novel SOCS1-specific antagonists can effectively restore anti-tumor immunity in this high-risk subgroup. Finally, we acknowledge several limitations. The low incidence of pediatric AML1::ETO with KIT mutations especially exon 17 and the requirement for fresh samples in HDS assays limited cohort size. Consequently, our findings are hypothesis-generating. In the absence of mechanistic validation, such as gain- and loss-of-function experiments or direct pathway activity measurements (e.g., pSTAT signaling), the role of the SOCS1/JAK–STAT axis remains correlative. Future studies incorporating larger multi-center cohorts and immune-competent in vivo models are necessary to validate these findings and further elucidate the interplay between KIT exon 17 mutations and SOCS1 upregulation. Conclusions In conclusion, KIT mutations, particularly in exon 17, confer a high-risk, chemoresistant phenotype to otherwise favorable pediatric AML1::ETO AML. Our preliminary experimental data suggest a potential chemoresistant profile in this subset, which might be mediated by SOCS1 upregulation and subsequent immune evasion. These findings highlight the critical need for routine KIT exon profiling and support the investigation of therapies targeting the SOCS1/JAK-STAT axis to improve outcomes for this specific patient subgroup. Supplementary Information Supplementary Material 1. Supplementary Material 1.