Profiling FLT3 Mutations in Mexican Acute Myeloid Leukemia Pediatric Patients: Impact on Overall Survival

Background: Acute myeloid leukemia (AML) is the second most frequent leukemia in childhood. The FLT3 gene participates in hematopoietic stem cell proliferation. FLT3 mutations are recurrent in AML and influence prognosis. In Mexican pediatric AML patients, FLT3 mutational profile, and their clinical impact have not been evaluated. Aim of the study: This study aimed to identify the profile of FLT3 mutations in pediatric patients with de novo AML and to assess their possible influence on overall survival (OS) and other clinical features. Methods: Massive parallel target sequencing of FLT3 was performed in 80 patients. Results: FLT3 mutations [internal tandem duplication (ITD) or tyrosine kinase domain (TKD)] were identified in 24% of them. OS was significantly lower in FLT3POS cases than in FLT3NEG (p = 0.03). The average OS for FLT3POS was 1.2 vs. 2.2 years in FLT3NEG. There were no significant differences in the children's sex, age, percentage of blasts in bone marrow aspirate, or white blood cell count in peripheral blood at diagnosis between both groups. No differences were identified stratifying by the mutational load (high > 0.4) or type of mutation. The negative effect of FLT3 mutations was also observed in patients with acute promyelocytic leukemia (APL). Conclusions: FLT3 mutational profile is described in Mexican pediatric AML patients for the first time. Mutated FLT3 negatively impacts the outcome of AML patients, even considering the APL group. The clinical benefit from treatment with tyrosine kinase inhibitors in the FLT3POS pediatric patients needs to be assessed in clinical trials. FLT3 testing may contribute to better risk stratification in our pediatric AML patients.


INTRODUCTION
Genomic investigations of acute myeloid leukemia (AML) have demonstrated that several genes are recurrently mutated, leading to the identification of new biomarkers. Genetic screening of molecular abnormalities plays a significant role in prognostic categorization and definition of treatment strategies in AML (1,2). FLT3 is a class III family receptor tyrosine kinase expressed in hematopoietic stem cells and has essential roles in cell survival and proliferation. FLT3 gene mutations are among the most frequently observed in AML, which occurs in approximately onethird of patients with de novo AML. They have been associated with the clinical prognosis, treatment response, and survival of patients (3).
Two types of FLT3 activating mutations have been identified in AML. Internal tandem duplications (ITDs) affect the juxtamembrane (JM) domain or tyrosine kinase domain 1 (TKD1) of the FLT3 receptor and point mutation preset in the TKD2 (4). Both types of mutations promote ligand-independent auto-phosphorylation and constitutive activation of the receptor. This unregulated activation impairs normal hematopoiesis and contributes to leukemogenesis (5).
ITD mutations are the most frequent and are considered drivers in AML. At diagnosis, the FLT3-ITD clone may possess a growth advantage. This clone becomes dominant at relapse through a clonal expansion process (6,7). Higher relapse rates and more reduced survival have been associated with FLT3-ITD in previous studies, largely depending on the allelic ratio (AR) of the mutations and the presence of mutations in other genes (8). The clinical impact of FLT3-TKD mutations is less clear (9).
FLT3 genetic testing, detecting both ITD and TKD mutations, is recommended by the National Comprehensive Cancer Network (10) and European Leukemia Net (3) for diagnostic workup of patients with AML, to identify those who may benefit from different targeted treatment options. FLT3 screening is not routinely performed in Mexico. Epidemiological data about the FLT3 mutational profile and its clinical impact in Mexican AML patients are minimal. Studies including a small number of de novo and secondary AML patients (children and adults) have reported a prevalence between 13 and 20% of FLT3 mutation using polymerase chain reaction (PCR)-based technologies (11).
The landscape of somatic mutations in pediatric AML was shown to be markedly different from adults, suggesting different biology and etiology for adult and childhood AML (12). These findings stress the importance of analyzing the mutational profile and evaluate the clinical impact of genetic alteration separately in both groups of patients. The purpose of this study was to assess the prevalence and the influence of FLT3 mutations on the outcome of pediatric AML de novo patients. AML is one of the main types of cancer in Mexican children; its incidence has a trend toward an increase in the last year, and mortality is higher than in developed countries despite that the same chemotherapy protocols are used (13).

Population
Eighty patients with de novo AML, coming from eight different public hospitals, were included. Patients were diagnosed between March 2010 and March 2018. AML diagnosis was performed at each institution using bone marrow aspirate and immunophenotype. The bone marrow samples analyzed in the present study were obtained at the time of the diagnosis, before treatment initiation, and submitted to the Mexican Inter-Institutional Group for Identifying Childhood Leukemia Causes in Mexico City.
A predefined set of data was collected from medical charts, including the child's sex, age, white blood cell (WBC) count in peripheral blood, and percentage of blasts in the bone marrow at diagnosis, French-American-British (FAB) classification, and treatment protocol.
The ethics and scientific review boards of the National Institute of Genomic Medicine, Mexico City, Mexico, approved this study (document number 28-2015-1). All human samples and clinical information were approved for its use in the present study. The child's parents signed the informed consent obtained in accordance with the Declaration of Helsinki.

DNA Extraction
DNA was extracted using Maxwell R

Next-Generation Sequencing
Sequencing of the exons 13-15 and 20 of the FLT3 gene was performed using the multigene panel "Myeloid solution" by SOPHiA Genetics (Sophia Genetics SA, Saint Sulpice, Switzerland). Library preparation and sequencing were performed according to the manufacturer's protocol. Capturebase enriched libraries were pooled (12 indexed libraries per pool) and sequenced on a MiSeq system using version 3 chemistry (Illumina, San Diego CA, USA). Sequencing data were analyzed with the Sophia DDM R software version 5.2.7.1 (Sophia Genetics SA, Saint Sulpice, Switzerland). Sequencing deep was higher than 1,000 × for all the target regions.
Variant fraction (VF) was computed by dividing the number of mutation-supporting reads by the total number of reads at the mutation site. AR was calculated by dividing the number of reads corresponding to the mutation by the number of reads according to the wild-type allele in the mutation site.

Statistical Analysis
The comparison of proportions between different groups was made using chi 2 or two-sided Fisher's exact test for dichotomic variables, when appropriate. A non-parametric Mann-Whitney U-test was used for continuous variables, p < 0.05 was considered as statistically significant. All calculations were performed using the SPSS software package, SPSS Version 21 (Chicago, IL, USA).
The Kaplan-Meier method (14) was used for assessing overall survival (OS). The log-rank test was used to evaluate differences between survival distributions with a 95% confidence interval (CI). The OS time was calculated from the day of diagnosis confirmation until either the last follow-up or patient's death from any cause. Patients who did not experience an event were censored at the time of the last follow-up. Those who did not attend a follow-up were censored at their date of the last known contact.

Demographic and Biological Characteristics of the Patients
The demographic, clinical, and main biological features of the patients are shown in Table 1. Fifty-five percent of the cases were male. The median age at diagnosis was 9.3 (range 0.4-17.5) years; the age at diagnosis was very similar in both genders. Acute promyelocytic leukemia (APL), M3, was the most prevalent subtype (36.3%), followed by the M2 (33.8%). M0, M5, and M6 were the less common subtypes. APL seems to occur more frequently in Hispanic populations than in other ethnic groups (15).
Almost 38% of the patients do not survive more than 1 year after diagnosis. The median OS in the study population was 1.95 years.
Patients were treated according to one of the following four protocols: BFM-1998, BFM-2001 (Berlin-Frankfurt-Münster Since chemotherapeutic drugs affect outcomes, we assess the impact of the treatment protocol in the OS (see Supplementary Figures 1A,B). No significant differences were found among patients treated with NOPHO-AML93, BFM-2001, or BFM-1998, although those patients receiving NOPHO-AML93 seem to have the best long-term outcome while BFM1998 the worst. Patients treated with these protocols were considered as a single group for comparison with the PETHEMA-APL-05 treated group. OS was significantly higher in APL patients in contrast to those children from other FAB subtypes and treated with any different therapeutic combination (Supplementary Figure 1B).

FLT3 Somatic Mutation Profile in Mexican Pediatric Acute Myeloid Leukemia Patients
Mutations in FLT3 were identified in 23.7% (19/80) of the patients, being more frequent in females (52.6%). The mutational profile is shown in Table 2. ITD mutations were present in 52.6% (10/19) of the FLT3 POS cases, with VFs ranging from 2.4 to 64.9%. The range of affected amino acids expanded from position 584 to 613. In one patient, in-frame deletion of amino acid 600 in the JM domain was observed, without any insertions or duplications of nucleotides. This patient was included in the ITD group since the mutation may affect the functioning of the JM domain. Missense or small deletion mutations in the TKD region were present in 47.4% (9/19) of the FLT3 POS patients with VFs ranging from 1.6 to 42.8%. Changes in position 835 of the TKD domain were the most frequent. Two different amino acid changes were observed: p.Asp835Tyr in three cases and p.Asp835His in two. Additionally, single nucleotide variations affecting positions 839 and 842 of TKD were observed in one patient each. Two inframe deletions affecting amino acid 836 and 837 were also found. A second FLT3 mutation, at a lower VF, was identified in two patients (M183 and M163) (Table 2, Figure 1).

Distribution of Demographic and Biological Features Between FLT3 Positive and Negative Patients
Age at diagnosis, sex, percentage of blasts in bone marrow, and FAB subtype distribution was compared between FLT3 POS and negative patients. The results are displayed in Table 3.
No significant differences were found. However, the median age at diagnosis was lower in the FLT3 POS group. When it was stratified by gender, FLT3 POS female patients were 2.1 years younger at diagnosis than FLT3 NEG females. In male FLT3 POS patients, the disease was diagnosed 0.4 years earlier than in the FLT3 NEG negative. None of these differences reached statistical significance.

Analysis of the Association Between FLT3 Mutation and Overall Survival
OS results, according to the FLT3 status, are shown in Figures 2A,B. The OS was significantly decreased among patients with FLT3 mutation (Figure 2A). The median of OS time was 1.2 (0.05-3.1) years in the FLT3 POS vs. 2.2 (0.05-7.6) years, in negative children ( Table 3). When ITD or TKD mutations were considered separated (Figure 2B), no differences were observed between survival curves from ITD or TKD positive patients vs. FLT3 NEG patients. The OS analysis was calculated with 79 patients due to the date for the last contact or death was not available in one patient.
Since the allelic burden of the FLT3 mutation has been shown to influence the outcome in adults, we stratified both FLT3-ITD and FLT3 POS (the type of mutation) patients according to their AR, using 0.4 as the cutoff value (16). The mutational burden did not influence OS, considering the FLT3-ITD group or the whole FLT3 POS group.
The impact of  group. The OS of APL-FLT3 POS was similar to the OS observed for other FAB subtypes grouped (p = 0.014).

DISCUSSION
Mutational profiling of patients with AML is part of routine diagnostic workups for both de novo and relapsed AML patients in centers from developed countries. In Mexico, the FLT3 mutational profile is not routinely performed in most public institutions; therefore, little epidemiological data exist about prevalence and heterogeneity of FLT3 mutations and their influence on clinical evolution in both AML adult and pediatric patients. This analysis of de novo AML cases represents the most extensive cohort of Mexican children and adolescents analyzed  (18).
More extensive analysis has been conducted in pediatric AML patients from the USA. In 2006, 19% of FLT3 POS cases (FLT3-ITD with 12% and FLT3-TKD with 6.7%) were detected in a group of 630 de novo AML pediatric patients, belonging to the Children's Cancer Group. The authors suggested that, in general, the frequency of mutated FLT3 is 15% lower in the pediatric population than in adults (19). However, in 2015, Tarlock et al. (20) reported a prevalence of 28% of FLT3 POS cases in a group of 799 pediatric AML patients (16.02% with ITDs, 4.6% with TKDs, and 7.6% with novel mutation).
Differences in sensitivity between the FLT3 detection methods used could be contributing to the lower prevalence of mutation observed in some of the previous studies compared with ours. Molecular methods based on PCR were employed in all of them. These methods can provide the length, but not the exact position or the sequence of the insertion, and have lower sensitivity than massive parallel sequencing technology. Mutations with meager allelic fraction could be missed by PCR-based methodologies. Interestingly, in this study, five patients had a VF lower than 4%. If these patients are removed from the prevalence calculation, the frequency of FLT3 mutation turns to be only 17.5%.
The prevalence of FLT3 mutation seems to be lower in adult AML Hispanic patients than in other groups, although technical limitations may have contributed to the observed differences (22). Ruiz Arguelles et al. reported 13% in 31 patients, while Arana Trejo et al. found 15% of FLT3 mutated in 21 patients; both studies were done in Mexican patients (11). A collaborative study, including patients from referral centers in Mexico and Colombia, reported only 20% of FLT3 POS patients. They analyzed 138 patients with de novo or secondary AML (11). This study also included the patients previously reported by Arana Trejo and Ruiz Arguelles (11). In AML adults from Argentina, Sánchez et al. (25) reported 20.6%, while in Brazil, 26.3% was found (26).
Although the frequency of FLT3 mutation was higher in the female group, no statistically significant difference in gender distribution was observed, similar to what has been found in most series (18). No association was seen between age at diagnosis, WBC, or blast percentage at bone marrow and the presence of FLT3 mutations.
ITD mutation was present in 52.6% (10/19) of the patients. The length of the alteration varies from 21 to 111 bp, being 21 bp, the only size observed in more than one patient. The size of the duplicated region ranges from 15 to 174 bp (19), but up to 400 bp has been previously reported (3,27).
In all cases, FLT3-ITD occurred in multiples of three nucleotides, preserving the reading frame of the transcript, as has been observed in other studies (3). Evaluation of the role of ITD size in clinical prognosis has opposite results. Some authors showed that increasing ITD size was associated with an adverse outcome (28), while others found no association (29). The number of patients in our study was too small to perform this analysis.
Furthermore, it has been suggested that ITD mutations always affect at least one amino acid residue from codon 591-597, with codon 597 being the most duplicated (30,31). We do not confirm this observation. In our series, ITD mutations expand from codon 584-613. ITDs involving codons 591-597 were present only in 50% of the patients with this type of mutation. Only four of them had the amino acid 597 included in the duplication. For the other 50% of the ITD patients, the mutation affected codons 600-613.
Most of the ITD mutations were located only at the JM domain (6/9 patients). However, in three cases, the mutation was found on the β1 sheet of the first kinase domain, corresponding to amino acids 610-615. In one patient (M191, p.Leu601_Glu611dup), the duplication alters both the JM domain and the first residues of the β1 sheet in TKD. In the other two patients (M126: p.Leu610_Glu611ins20 and M169: p.Phe612_Gly613ins37), the integration occurred exclusively in the region of the β1 sheet of the first kinase domain. ITDs affecting the β1 sheet represent 22.2% of the ITD patients in this study, very similar to the frequency found in other series. No ITDs affecting different regions of the TKD were observed, although they have been described in a proportion of ITD patients (4).
In one patient, the in-frame loss of three nucleotides, producing Asp600 deletion in the JM domain, was observed. That mutation had a low VF. It does not correspond to the classic insertion/duplication occurring in the JM domain and has not been previously reported. It has been suggested that even a minimal mutation would be able to disrupt the intrinsic negative regulatory effects of the JM domain in preventing dimerization without ligand stimulation and would be sufficient to lead to auto-phosphorylation in cis (32). It could be hypothesized that this small deletion originates gain of function of the FLT3 receptor, similar to ITD involving a more significant number of amino acid residues.
Revealing the heterogeneity of ITD mutations seems to be important in the clinical setting since it could be one of the factors leading to variability in the treatment response of the FLT3-ITD AML patients. Location of the ITD in FLT3 influences the sensitivity to tyrosine kinase inhibitors as well as disease progression in mice (33). TKD-ITDs have been associated with a worse survival prognosis and to chemotherapy resistance in comparison with those with JM-ITD (30). In this study, only two patients with ITD-TKD were found; both had a high mutational burden. One of them had the highest mutational burden observed in our study (1.85 in patients M169). Marhäll et al. (34) showed that TKD-ITD and JM-ITD mutations display a similar oncogenic potential and that it is higher than the one from the D835Y point mutation located in the activation loop of the TKD.
FLT3-TKD mutations represent 47.4% (9/19) of the FLT3 POS cases, and 11.25% (9/80) of the total of patients, a higher prevalence than the 7.7% reported for AML adult patients (35). Six different mutations were observed; two of them were not listed in the COSMIC database (as of June 2020): p.Tyr842Ser, and, p.Met837del. Mutations within the activation loop of the second TKD were the most frequent (five of nine TKD cases). Asp835Tyr, considered the most common substitution in position 835, was found in only three patients, while Asp835His was identified in the other two cases.
FLT3-Asp835 mutations lead to constitutive activation of FLT3 (36) and have been reported in 7-14% of AML (3), 3% of myelodysplastic syndromes, and 3% of acute lymphoblastic leukemia cases (37). Other reported substitutions, including Asp835Val, Asp835Glu, and Asp835Asn, were not found in our patients. The activating mutation p.Ile836del, identified in one patient, was described for the first time by Thiede et al. (35) in 13/979 AML adult patients.
The mutation p.Tyr842Ser was identified in one patient with VF = 26%. p.Tyr842Cys affecting the same codon has been associated with resistance to midostaurin, sunitinib, sorafenib, lestaurtinib, KW2449, and AGS324 (38). The main challenge with FLT3 inhibitors and other target therapies is to overcome primary and secondary resistance. The possibility that Tyr842Ser could be related to primary resistance should be considered in the clinical setting. Since none of our patients were treated with FLT3 inhibitors, the possible role of this mutation in response cannot be evaluated.
In an extensive series of pediatric AML patients, Bolouri et al. (12) reported mutations not found in adults with AML. New  FLT3 mutations affecting residues 451, 444, 491, 676, 680, and 941 were described, some of them with functional implications and influencing response to conventional therapy (20). If these mutations exist in our patients, they could not be detected due to target sequencing strategy limitations. Only mutations located between amino acids 534-647 and amino acids 807-847 could be detected. A full gene sequencing strategy is the best option to explore somatic mutations in FLT3 and other genes too.
Two patients (22.2% of the FLT3 POS group) presented a second mutation at a lower VF, suggesting the presence of additional clones at diagnosis. A similar observation was reported by Kottaridis et al. (39) who found more than one mutation in 23% of FLT3-ITD AML patients, with up to five different FLT3-ITD clones of various sizes, insertion sites, and ARs identified. A worse OS has been observed in patients with more than one clone (39).

Analysis of the Prognosis Impact of FLT3 Mutation
OS was analyzed considering ITD and TKD mutations separated; both groups of patients showed lower survival than did the FLT3 NEG group (Figure 2). The influence of TKD or ITD on prognosis was very similar; no significant difference in OS between both groups of patients was observed.
The FLT3-ITD mutations have been previously related to a poor outcome in both adults (40,41) and pediatric patients (42,43). Wu et al. (44) performed a meta-analysis, including 1661 pediatric patients with AML. Patients with FLT3-ITD mutation had an inferior OS [hazard ratio (HR) = 2.19 (1.60-3.01), p < 0.001] in comparison with patients without FLT3 mutation. The impact of TKD mutations was not analyzed in this paper. The author suggested that FLT3-ITDs produce a significantly negative prognostic effect in pediatric patients with AML.
There is no clear explanation of why FLT3-ITD could have such a negative impact. FLT3-ITD mutations contribute to increased production of reactive oxygen species (45), which leads to DNA oxidative damage, increases DNA double-strand breaks, and mistakes in the repair mechanisms starting a cycle of genomic instability. The existence of genomic instability leading to a cytogenetic evolution in AML is supported by the acquisition of new structural chromosomal abnormalities by leukemic cells between diagnosis and relapse. McCormick et al. (46) showed that cytogenetic evolution was more frequent among FLT3-ITD AML patients. Ten out of 14 FLT3-ITD AML patients acquired new cytogenetic and structural chromosomal abnormalities, compared with seven out of 21 FLT3 wildtype AML patients. This observation has been confirmed by others analyzing FLT3-ITD AML patients at relapse (47). Other genomic features supporting genomic instability in FLT3-ITD patients have been shown, including microsatellite instability (48,49), shorter telomeres (50), and high frequency of somatic mutations (51). We speculate that acquisition of additional poor-prognosis mutation is favored by the presence of genomic instability in FLT3-ITD patients and might be one of the factors contributing to the detrimental effect of FLT3 mutations on survival (52).
Genomic instability is a hallmark of cancer, and high rates of genomic instability are present in other myeloid malignancies containing activated tyrosine kinase pathways, such as BCR/ABL in chronic myeloid leukemia, JAK2 in myeloproliferative neoplasias, and RAS mutations in myelodysplastic syndromes (45).
The effect of TKD mutations on prognosis has been mainly evaluated in adult patients and is considered controversial (3). Mead et al. (53) showed that patients with FLT3-TKD have a significantly more favorable prognosis than FLT3-ITD AML patients. In contrast, Whitman et al. (54) found that FLT3 D835/I836 mutations were associated with poor diseasefree survival and a distinct gene-expression signature among younger adults. Our results suggest that both FLT3-TKD and ITD mutations have a similar negative effect on OS in pediatric patients. Still, additional factors, like cytogenetic background, mutations in other genes, and treatment, may be contributing to the negative outcome in the TKD group. For example, the number of patients no-APL is significantly higher in TKD than in the ITD group (77.8% TKD vs. 30% ITD; p Fisher = 0.046). Patients with other FAB subtypes had OS significantly lower than the APL group, treated with the PETHEMA-APL-05 (see Supplementary Figure 1B).
When the OS was considered, regardless of the subtype of FLT3 mutation, patients with FLT3 mutation had significantly lower OS than the FLT3 NEG patients, supporting that constitutive activation of the FLT3 receptor, for any subtype of mutation, predisposes to high-risk disease. The negative effect of an FLT3 mutation was additionally supported by the analysis in the APL group of patients. All APL patients were treated with the same protocol. However, APL-FLT3 POS patients had significantly worse OS that M3-FLT3 NEG patients. Other studies have shown an adverse effect of mutated FLT3 in APL patients. Analyses of 205 adults and children found a significantly higher risk of death among APL patients with FLT3-ITD than in those without (HR = 11.74; 95% CI = 1.03-134.5), a reduced OS was also observed (15). Additionally, a higher relapse rate and a more reduced postrelapse survival have also been reported in the APL FLT3-ITD as compared with patients with APL FLT3 wild-type (55,56).
Both subtypes of FLT3 mutation yield proteins that spontaneously dimerize, bypassing ligand-mediated activation. When they were transfected into murine cell lines, factorindependent growth originated (57). The analysis of 979 AML adult patients showed that both aberrations are also associated with similar clinical features: higher WBC and higher numbers of bone marrow blasts (35).
The impact of constitutive FLT3 activation has been evaluated in several murine models. FLT3-ITD expression originated a myeloproliferative disorder having different phenotypes and severity depending on the murine model. In a bone marrow transplant model, a fatal myeloproliferative neoplasm occurred (57), while in a knock-in FLT3-ITD model, a less severe human chronic myelomonocytic leukemia like disease resulted. FLT3 constitutive activation alone does not produce a full AML genotype; additional cooperating genetic alterations are needed in human and mouse for AML development (58).
FLT3-ITD is not expressed to the same level in all patients; differences in expression measured using the FLT3-ITD to wildtype AR also impact prognosis (3). It has been retrospectively shown that patients with a higher FLT3-ITD mutant-to-wildtype ratio have significantly worse outcomes than patients with a lower ratio, being 0.4 the cut-point in children and 0.5 in adults (59). In our ITD patients, the allelic burden ranged from 0.02 to 1.85; it was higher than 0.4 in 66.7% (6/9) of them. This study did not find that the mutational burden influences OS in the FLT3-ITD group or the whole FLT3 POS group. A significant limitation is the small sample size impairing the statistical power of the analysis. The role of the FLT3 AR on the patient's outcome needs further clarification in prospective analysis with larger number of patients.
It has been suggested that when the FLT3-ITD AR is low, the presence of NPM1 mutation may mitigate the adverse prognostic effects of the FLT3-ITD mutation (60). However, all FLT3 POS patients had wild-type NPM1. Mutation in this gene was identified only in one of the cases (unpublished data), supporting the observation that this alteration is more common in adults than in pediatric AML patients (8).
Several limitations for this study must be considered: the results were not based on a randomized controlled trial but a retrospective analysis of a heterogeneously treated group of patients coming from eight different institutions. Patients were treated according to four different protocols that had different effects on the outcome. Besides, cytogenetic/fluorescence in situ hybridization (FISH) characterization, considered an independent prognostic factor in AML, was not performed in these patients. This study was not able to assess the potential effects of chromosomal aberration and mutations in other genes and its possible interaction with FLT3 alterations. Since a full FLT3 gene sequencing was not performed, mutation occurring outside exons 13-15 and exon 20 were not detected. Therefore, the frequency of FLT3 mutation in Mexican pediatric AML patients may be even higher than the one obtained in the present work. This is the first study evaluating the FLT3 mutational profile in pediatric de novo AML Mexican patients and its impact on the prognostic. The results suggest that FLT3 mutation could be considered as an independent biomarker for high-risk disease in pediatric AML patients and highlights the need for testing FLT3 routinely in the clinical setting. A risk-adapted management decision may help to improve the clinical outcome in AML patients, similar to what has been achieved in adult AML patients.
The trend toward better response rates has seen in patients who received chemotherapy as first-line therapy in combination with FLT3 inhibitors compared with those who did not receive it (3). Stone et al. (61) reported the results of an international randomized phase III study, including adult AML patients. Induction therapy (daunorubicin and cytarabine) and consolidation therapy (high-dose cytarabine), plus either midostaurin or placebo, revealed improved 4-year survival in the midostaurin arm compared with the placebo. These results were independent of allelic burden (high vs. low) or the type of FLT3 mutation present (ITD or TKD). The benefit of midostaurin was observed among patients who underwent transplantation during the first remission but not among those who underwent transplantation at a later time (61). Even after intensive chemotherapy and stem cell transplant, FLT3-ITD is an adverse prognostic factor that remains prognostically relevant (3).

CONCLUSIONS
A mutant FLT3 is present in approximately one out of four pediatric AML patients in Mexico, and it is related to adverse outcomes. OS was impaired even in the APL subgroup that nowadays has higher complete remission and cure rates than other FAB subgroups due to more efficient therapeutic combinations.
All patients with newly diagnosed AML should have the FLT3 tested using whole-genome sequencing approach better than a specific exon-target enrichment method to be able to detect some clinically relevant children with particular mutations. Earlier identification of FLT3 mutational status will allow identification of patients with a more aggressive disease that could benefit from modification in the treatment strategy, including entering clinical trials, receiving focus therapy, or be considered for bone marrow transplant.
Increasing awareness about the importance and clinical utility of FLT3 testing in AML is the first step to make FLT3 testing part of routine diagnostic workups for these patients in Mexico.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by National Institute of Genomic Medicine, Mexico City, Mexico (number of approval 28-2015-1). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AUTHOR CONTRIBUTIONS
CM obtained the results, analyzed data, and drafted and approved the manuscript. KC, LLF, MJ, AMR, and BV obtained and validated the results and approved the manuscript. HF validated the results, analyzed statistical data, and approved the manuscript. JN, EJ, VB, JT, JF, JAM, MM, AMS, LE, JP, RE, LVF, RA, MP, OS, HR, AR, MD, EG, PG, and JMM collected clinical data, cared for patients, and approved the manuscript. MD, EG, and ER, validated the results and approved the manuscript. CA carried out the general oversight and the acquisition of funds, designed the study, obtained and validated the results, analyzed the data, and drafted and approved the manuscript. All authors contributed to the article and approved the submitted version.

ACKNOWLEDGMENTS
We are grateful to all patients, their families, and the clinical and technical personnel who participated in this study.