Advances in the First Line Treatment of Pediatric Acute Myeloid Leukemia in the Polish Pediatric Leukemia and Lymphoma Study Group from 1983 to 2019

Simple Summary We retrospectively analyzed the results of the five consecutive treatment protocols for pediatric acute myeloid leukemia (AML) used in Poland from 1983 to 2019 (excluding promyelocytic, secondary, biphenotypic, and Down syndrome AML). The study included 899 children. The probability of three-year overall, event-free, and relapse-free survival increased from 0.34 ± 0.03 to 0.75 ± 0.05, 0.31 ± 0.03 to 0.67 ± 0.05, and 0.52 ± 0.03 to 0.78 ± 0.05, respectively. A systematic reduction of early deaths and deaths in remission was achieved, while the percentage of relapses decreased only in the last therapeutic period. Surprisingly good results were obtained in the group of patients with unfavorable genetic abnormalities like KMT2A-MLLT10/t(10;11)(p12;q23) and DEK-NUP214/t(6;9)(p23;q24) who were treated in the AML-BFM 2012 Registry, while an unsatisfactory outcome was found in patients with FLT3-ITD. The use of standardized therapeutic protocols with the successive consideration of genetic prognostic factors and advances in supportive care led to a significant improvement in AML treatment outcomes over the last 40 years. Abstract Background: From 1983, standardized therapeutic protocols for pediatric acute myeloid leukemia (AML) based on the BFM group experience were introduced in Poland. We retrospectively analyzed the results of pediatric AML treatment in Poland from 1983 to 2019 (excluding promyelocytic, therapy-related, biphenotypic, and Down syndrome AML). Methods: The study included 899 children suffering from AML treated with the following: AML-PPPLBC 83 (1983–1993, n = 187), AML-PPGLBC 94 (1994–1997, n = 74), AML-PPGLBC 98 (1998–2004, n = 151), AML-BFM 2004 Interim (2004–2015, n = 356), and AML-BFM 2012 (2015–2019, n = 131). Results: The probability of three-year overall survival was 0.34 ± 0.03, 0.37 ± 0.05, 0.54 ± 0.04, 0.67 ± 0.03, and 0.75 ± 0.05; event-free survival was 0.31 ± 0.03, 0.34 ± 0.05, 0.44 ± 0.04, 0.53 ± 0.03, and 0.67 ± 0.05; and relapse-free survival was 0.52 ± 0.03, 0.65 ± 0.05, 0.58 ± 0.04, 0.66 ± 0.03, and 0.78 ± 0.05, respectively, in the subsequent periods. A systematic reduction of early deaths and deaths in remission was achieved, while the percentage of relapses decreased only in the last therapeutic period. Surprisingly good results were obtained in the group of patients treated with AML-BFM 2012 with unfavorable genetic abnormalities like KMT2A-MLLT10/t(10;11)(p12;q23) and DEK-NUP214/t(6;9)(p23;q24), while unsatisfactory outcomes were found in the patients with FLT3-ITD. Conclusions: The use of standardized, systematically modified therapeutic protocols, with the successive consideration of genetic prognostic factors, and advances in supportive care led to a significant improvement in AML treatment outcomes over the last 40 years.


Introduction
Acute myeloid leukemia (AML) is a heterogeneous disease with a clonal expansion of myeloid progenitors in the bone marrow and peripheral blood. It comprises about 15% of all leukemias in children. The treatment results in pediatric AML have improved significantly over the last 40 years, and the current overall survival is about 70% [1,2]. A history of unified therapeutic pediatric AML protocols in Poland began in 1983. The Polish Pediatric Leukemia/Lymphoma Study Group (PPLLSG) introduced protocol AML-PPLLSG 83, a modified version of Berlin-Frankfurt-Munster (BFM)-AML 83 protocol. The main aim of PPLLSG was to develop diagnostic and treatment standards, and to trial the utility of intensive protocols in Polish conditions. The protocol allowed for increasing the five-year EFS to 32% compared with less than 15% before 1983 [3,4]. However, the results were still significantly worse than the original BFM results [3,5]. The main cause of the difference was a high percentage of early deaths, resulting in a significantly lower remission rate. Insufficient supportive care and toxicities were the main restriction in the introduction of high doses of cytarabine in the consecutive protocols. In the new protocol (AML-PPLLSG 94) stratification into risk groups was implemented. This caused a slight, but not significant, increase in the event-free survival (EFS) rate (from 32 to 36%) [3,4]. However, the ID-ARAC courses were not effective at producing morphological remissions in late responders and the remission rate did not increase [3]. In the consecutive protocol, AML-PPLLSG 98, a more accurate stratification system was implemented and a new drug, idarubicine, was introduced instead of daunorubicine, for the induction, intensification, and consolidation. Better initial responses were observed in the SR group; however, unsatisfactory treatment outcomes were found in the HR group [3,4].
In 2005, the AML-BFM 2004 Interim protocol was introduced in PPLLSG centers. The stratification to the risk groups was based on the FAB classification, genetic abnormalities, and response to the treatment. Since that time, the central verification of the cell morphology, immunophenotyping, and cytogenetic examination has been available. Since 2006, molecular studies (fusion genes AML1-ETO, PML-RARA, CBFβ-MYH11, MLL-AF4, MLL-AF9, and MLL-ENL, as well as FLT3-ITD and WT1 overexpression) have been performed in the central laboratory in Krakow as the standard diagnostic procedure in all pediatric patients with AML. Allogeneic hematopoietic stem cell transplantation (allo-HSCT) from HLA-matched sibling donors was introduced for all patients of the HR group in the first morphological remission, and allo-HSCT from an HLA-matched unrelated donor for children with persistent blasts after the second induction (HAM) or still in aplasia 4 weeks after the second induction. A further improvement of the treatment results, an increase in the proportion of achieved remission, and a reduction of early deaths and deaths in remission was observed; however, the proportion of relapses was still high.
In 2015, a new treatment protocol AML-BFM 2012 was introduced. Stratification to the risk groups was based on cytogenetic and molecular genetics changes. The molecular diagnostic is extended by the HemaVision kit and an analysis of the additional gene mutations (NPM1 and CEBPα). Three risk groups were defined. Cladribine was added for the intermediate and high-risk group in the third chemotherapy course. Hematopoietic stem cell transplantation was indicated in all patients in the high-risk group. A further improvement of the treatment results was achieved.
In parallel with the therapeutic protocol improvements, supportive care also developed, including antifungal prophylaxis, broad spectrum antibiotic in case of neutropenic fever, and guidelines for emergencies like hyperleukocytosis and tumor lysis syndrome. This is retrospective analysis of the results of the five consecutive treatment protocols for pediatric AML used in Poland from 1983 to 2019.

Materials and Methods
From 1983 to June 2019, there were 1225 patients from 0 to 18 years old with novo AML diagnosed in Poland. We excluded patients with secondary AML (40), acute promyelocytic leukemia (98), biphenotypic leukemia (23), and patients with Down syndrome myeloid leukemia (68). In addition, 17 patients died before the beginning of the treatment, 15 patients were pretreated with another protocol because of wrong initial diagnosis (acute lymphoblastic leukemia, neuroblastoma), 5 patients could not follow the treatment protocol because of coexisting morbidities, and 16 patients were excluded because of the lack of data. Finally, 899 patients were enrolled in the retrospective analysis. The data were collected through the Polish Acute Myeloid Leukemia Registry.
The patient characteristics are presented in the Table 1. Details of the treatment are presented in Figure 1 and Table 2. Table 3 contains stratification to the risk groups in the periods II−V. Indications for HSCT and the number of patients who underwent the procedure are shown in Table 4.  t(4;11), t(5;11), t(6;11), t(10;11), t(6;9), t(7;12), der12p, isolated monosomy 7, t(9;22), FLT3-ITD-WT1mut, complex karyotype SRG-standard risk group; IRG-intermediate risk group; HRG-high risk group; BM-bone marrow; a In case of FLT3-ITDreclassification to HRG; b In case of BM blasts >5% on day 15 or blastic reconstitution between days 15 and 28-reclassification to HRG; c In case of BM blasts >20% on the days 21−28-reclassification to IRG; d In case of BM blast > 20% on the days 21−28 or >5% on the days 42−56-reclassification to HRG. CSN prophylaxis and therapy are presented in Table 2.

Diagnostics
The diagnosis was based on bone marrow (BM) examination, including morphology with FAB classification and immunophenotyping. Cytogenetic analyses (both classical and fluorescence in situ hybridization (FISH)) performed in local treatment centers were mandatory from 1998. The results of the karyotype were eligible in 449 patients, including 67 (44%), 267 (75%), and 115 (88%) from periods III, IV, and V, respectively. Since 2005 (period IV and V), central verification of the cell morphology and immunophenotyping examination has been introduced. Moreover, since 2006, molecular studies (fusion genes AML1-ETO, PML-RARA, CBFβ-MYH11, KMT2A-MLLT1, KMT2A-MLLT3, and KMT2A-AFF1) have also been performed in the central Genetic Laboratory of the University Children's Hospital of Krakow. The results are available for 292 patients from period IV (82%). Since 2015 (period V), the molecular diagnostics has been extended by the HemaVision kit and through analysis of additional gene mutations (NPM1 and CEBPα). These results are available for 125 patients for period V (95%). Analysis of the FLT3-ITD was performed from 2005, and the results for 410 patients are available (278 from period IV (78%) and 129 from period V (100%)). HSCT-hematopoietic stem cell transplantation; HRG-high risk group; SRG-standard risk group; MSD-matched sibling donor; MUDmatched unrelated donor; MD-mismatched donor; Haplo-HSCT from haploidentical donor; OS-overall survival; CR-complete morphological remission; Allo-allogeneic; Auto-autologous; nd-no data.
The data concerning the cytogenetics and molecular genetics of the patients from periods IV and V are presented in Table 5.

Definitions
Complete morphological remission is defined as less than 5% of blasts in BM demonstrating normal or only slightly decreased cellularity, with signs of regeneration of normal hematopoiesis, regeneration of normal cell production in peripheral blood (≥1000/µL leukocytes, ≥500/µL neutrophil granulocytes and ≥50,000/µL platelets), lack of blasts in peripheral blood, and the disappearance of any extramedullary sites. Minimal residual disease was assessed only in the last period as a part of a separate study; these results did not influence the therapeutic decision and were not analyzed in that paper.
Treatment failure included early death (death during the first 42 days from the start of the treatment); failure to achieve morphological remission on the first line therapy (no fulfillment of CR criteria by the end of intensification; leukemic blasts ≥5% after the second induction; aplasia ≥6 weeks after the beginning of the second induction) and relapse (the reappearance of leukemic blasts in the peripheral blood; re-infiltration of BM with >5% distinct blasts not attributable to any other cause; leukemic infiltration elsewhere following CR; or partial remission lasting at least 4 weeks).
Positive (% of available) 30 (10.8) 11 (8.5) Statistical analyses were performed with STATISTICA, version 13, StatSoft Inc. Probabilities of survival were calculated using the Kaplan−Meier method. Overall survival (OS) was defined as the time from the date of diagnosis to the date of death from any cause, or from the last follow-up. Event-free survival (EFS) was calculated as the time from diagnosis to the first event (relapse, death of any cause, failure to achieve remission, or secondary malignancy) or last follow-up. Failure to achieve morphological remission was considered an event on day 0. Relapse-free survival (RFS) was calculated as the time from the first remission to the relapse. Probabilities of survival were presented as decimal fractions with standard deviations. The sub-groups were compared with a log-rank test. The relative risks were estimated using a Cox proportional hazard model. The results are presented as a hazard ratio (HR) with a 95% confidence interval (CI). All variables having a p value <0.05 in the univariate analysis were included in a multivariate analysis.
The study was conducted according to the guidelines of the Declaration of Helsinki, and was approved by the Ethics Committee of Jagiellonian University (protocol code: 122.6120.17.2015, date of approval: 29 January 2015). Informed consent was obtained from all patients or guardians when entering each therapeutic protocol.

Treatment Results
The percentage of achieved remissions increased from 72% in period I to 87% in periods IV and V. First-line non-responders comprised 6%, 18%, 10%, 8%, and 7%, respectively. A reduction of early deaths (up to the 42nd day of treatment) was noted, and the percentage decreased systematically in periods I−V from 22% to 5%. Among the patients who achieved remission, the percentage of deaths from toxicities in the following periods decreased from 16% to 1.6%. The rates of relapse were similar in periods I−IV (about 30%) and decreased in the last period to 17% (median follow-up in period V of 37 months; Table 1). The results are presented in Figure 2 and Table 6. riods IV and V. First-line non-responders comprised 6%, 18%, 10%, 8%, and 7%, respectively. A reduction of early deaths (up to the 42nd day of treatment) was noted, and the percentage decreased systematically in periods I−V from 22% to 5%. Among the patients who achieved remission, the percentage of deaths from toxicities in the following periods decreased from 16% to 1.6%. The rates of relapse were similar in periods I−IV (about 30%) and decreased in the last period to 17% (median follow-up in period V of 37 months; Table  1). The results are presented in Figure 2 and Table 6.  In the subsequent periods, survival improved significantly. The probability of threeyear OS increased from 0.34 ± 0.03 to 0.75 ± 0.05, EFS from 0.31 ± 0.03 to 0.67 ± 0.05, and RFS from 0.52 ± 0.03 to 0.78 ± 0.05. The survival curves are presented in Figure 3, and the probabilities of survival are in Table 6.   In the subsequent periods, survival improved significantly. The probability of threeyear OS increased from 0.34 ± 0.03 to 0.75 ± 0.05, EFS from 0.31 ± 0.03 to 0.67 ± 0.05, and RFS from 0.52 ± 0.03 to 0.78 ± 0.05. The survival curves are presented in Figure 3, and the probabilities of survival are in Table 6.

Risk Groups
The outcome within the risk groups was analyzed in the last two periods. A comparison of the survival between risk groups was performed separately for each period, as the definition of the risk groups differed across the therapeutic protocols.
According to the AML-BFM 2004 protocol (period IV), two risk groups were distinguished. There were 71 patients (19.9%) in the SRG and 284 children (79.8%) in the HRG (no data in 1 patient). The probability of five-year OS was significantly higher in SRG (0.74 ± 0.04) compared with HRG (0.62 ± 0.03; p = 0.03178). The difference in the probability of EFS was almost significant between SRG and HRG (0.56 ± 0.04 vs 0.47 ± 0.03; p = 0.0567). No significant differences were found in the probability of RFS (SRG 0.59 ± 0.04, HRG 0.61 ± 0.03). The survival curves are shown in Figure S1.

Risk Groups
The outcome within the risk groups was analyzed in the last two periods. A comparison of the survival between risk groups was performed separately for each period, as the definition of the risk groups differed across the therapeutic protocols.
According to the AML-BFM 2004 protocol (period IV), two risk groups were distinguished. There were 71 patients (19.9%) in the SRG and 284 children (79.8%) in the HRG (no data in 1 patient). The probability of five-year OS was significantly higher in SRG (0.74 ± 0.04) compared with HRG (0.62 ± 0.03; p = 0.03178). The difference in the probability of EFS was almost significant between SRG and HRG (0.56 ± 0.04 vs. 0.47 ± 0.03; p = 0.0567). No significant differences were found in the probability of RFS (SRG 0.59 ± 0.04, HRG 0.61 ± 0.03). The survival curves are shown in Figure S1.
In period V, three risk groups were defined. There were 25 children in SRG (18.5%), 52 patients in IRG (40%), and 52 in HGR (40%); no data were available concerning two patients. The probabilities of five-year OS were 0.92 ± 0.08, 0.81 ± 0.06, and 0.64 ± 0.06 in SRG, IRG, and HRG, respectively. There was a significant difference in OS between SRG and HRG (p = 0.01296). The probability of five-year EFS was significantly higher in SRG compared with HRG (0.83 ± 0.08 vs. 0.53 ± 0.06; p = 0.01925), with no difference in comparison to IRG (0.69 ± 0.06). No significant difference between the risk groups was noted in terms of RFS (0.83 ± 0.9, 0.71 ± 0.06, and 0.74 ± 0.07 in SRG, IRG, and HRG, respectively). Survival curves are presented in Figure S2 Figure S3). There were no significant differences in OS, EFS, and RFS in the FLT3-ITD positive patients, depending on the treatment protocol.

Fusion Genes
The results of CBFB-MYH11 and RUNX1-RUNX1T1 were available for most patients treated with AML-BFM 2004 (period IV) and AML-BFM 2012 (period V). There were 24 patients (6%) with the CBFB-MYH11 fusion gene (12 in period IV and 12 in period V). The probabilities of five-year OS, EFS, and RFS were 0.82 ± 0.08, 0.65 ± 0.11, and 0.74 ± 0.11, respectively. The survival curves are presented in Figure S4. No resistant disease was noted in that group of patients. In period IV, there were four relapses and three patients died (all because of toxicities, including one after HSCT in the second remission). In period V, there were no relapses, and one patient died in remission because of severe infection. The RUNX1-RUNX1T1 fusion gene was identified in 66 patients (16%) in the analyzed group (51 in the period IV and 11 in the period V). Only one patient (treated with AML-BFM 2004) did not achieve morphological remission and died in the course of disease progression. Twenty children relapsed (16/51 (31%) in period IV and 4/16 (25%) in period V), including three patients with two relapses (one of them died because of the disease progression). In total, 13 children died-2 of progression (1 non-responder and 1 after second relapse) and 11 of toxicities (4 in the I CR and 7 in the II CR, including 4 after HSCT). All of the deaths occurred in period IV. The probabilities of five-year OS, EFS, and RFS were 0.79 ± 0.05, 0.62 ± 0.06, and 0.68 ± 0.06, respectively. Survival curves are shown in Figure S5. Probability of OS was almost significantly higher in period V compared with period IV (1.0 vs. 0.74, p = 0.05239), and no significant differences were found in the probability of EFS and RFS between periods IV and V (0.61 ± 0.07 vs. 0.71 ± 0.1 and 0.67 ± 0.08 vs. 0.71 ± 0.12; Figure S6).
An analysis of survival depending on the other fusions genes was performed only for the patients from period V. In that group, the HemaVision test including 28 fusion genes was used. The results were available for 125 children (95%). Survival curves were done for the fusion genes found in at least five patients, including: CBFB-MYH11 (12 children), RUNX1-RUNX1T1 (16)

Hematopoietic Stem Cell Transplantation
In the whole analyzed cohort, 173 patients received HSCT in the first CR. As the number of transplantations in the first two periods was low and the data concerning HSCT from that time were lacking, the next analysis considered the children from periods III−V. During that time, HSCT was performed in 157 patients, including 20 in period III, 100 in period IV, and 37 in period V. There were 10 auto-HSCT and 147 allo-HSCT (Table  4).
An analysis of the survival was performed for the patients with event-free survival longer than 4 months (HSCT was performed after at least four chemotherapy cycles, which means after at least 4 months from diagnosis). There was a significant difference in RFS between patients with and without HSCT (five-year RFS 0.74 ± 0.04 vs. 0.62 ± 0.03, p = 0.02252), while there were no differences in OS and EFS (five-year OS 0.79 ± 0.03 vs. 0.74

Hematopoietic Stem Cell Transplantation
In the whole analyzed cohort, 173 patients received HSCT in the first CR. As the number of transplantations in the first two periods was low and the data concerning HSCT from that time were lacking, the next analysis considered the children from periods III−V. During that time, HSCT was performed in 157 patients, including 20 in period III, 100 in period IV, and 37 in period V. There were 10 auto-HSCT and 147 allo-HSCT (Table 4).
An analysis of the survival was performed for the patients with event-free survival longer than 4 months (HSCT was performed after at least four chemotherapy cycles, which means after at least 4 months from diagnosis). There was a significant difference in RFS between patients with and without HSCT (five-year RFS 0.74 ± 0.04 vs. 0.62 ± 0.03, p = 0.02252), while there were no differences in OS and EFS (five-year OS 0.79 ± 0.03 vs. 0.74 ± 0.02, p = 0.20251; EFS 0.69 ± 0.03 vs. 0.60 ± 0.02, p = 0.14179). Survival curves are shown in Figure 5.

Discussion
The treatment results in childhood AML in Poland improved significantly during the last 40 years. Probability of OS, EFS, and RSF increased from 0.31 ± 0.03 to 0.75 ± 0.05, from 0.30 ± 0.03 to 0.65 ± 0.05, and from 0.51 ± 0.03 to 0.75 ± 0.05, respectively. At the beginning of the history of unified AML treatment protocols in Poland, a high number of early deaths was observed, leading to a lower percentage of morphological remission. In the protocol AML-PPLLSG 94 with intermediate doses of cytarabine, the percentage of early deaths was reduced; however, the non-responders rate increased. Improvement of the supportive care enabled further intensification of the treatment. Introducing idarubicine in protocol AML-PPLLSG 98 and high-dose cytarabine in AML-BFM 2004 led to a significant decrease of non-responders compared to earlier therapy, and a significant increase of morphological remission. This was consistent with the reports of the BFM Study Group [6,7]. Despite the improvement in the treatment outcome in periods I-IV, the percentage of relapses remained unsatisfactory high. A significant reduction of relapses was observed in the last period (AML-BFM 2012 Registry); however, the follow-up was much shorter (median 37 months) than in previous periods.
In addition to therapeutic protocol improvement, supportive care developed, including antifungal prophylaxis, broad spectrum antibiotic for neutropenic fever, and experience in management with emergencies like hyperleukocytosis or tumor lysis syndrome. All of this contributed to a reduction in early deaths and deaths in remission, and finally to an increase in overall survival. The significant role of supportive care in the improvement of the AML treatment outcome was reported by other authors [2,[8][9][10][11].
Identification of the prognostic factors enabled stratification to the risk groups. At the beginning, it was based on the treatment response and FAB classification. Further

Discussion
The treatment results in childhood AML in Poland improved significantly during the last 40 years. Probability of OS, EFS, and RSF increased from 0.31 ± 0.03 to 0.75 ± 0.05, from 0.30 ± 0.03 to 0.65 ± 0.05, and from 0.51 ± 0.03 to 0.75 ± 0.05, respectively. At the beginning of the history of unified AML treatment protocols in Poland, a high number of early deaths was observed, leading to a lower percentage of morphological remission. In the protocol AML-PPLLSG 94 with intermediate doses of cytarabine, the percentage of early deaths was reduced; however, the non-responders rate increased. Improvement of the supportive care enabled further intensification of the treatment. Introducing idarubicine in protocol AML-PPLLSG 98 and high-dose cytarabine in AML-BFM 2004 led to a significant decrease of non-responders compared to earlier therapy, and a significant increase of morphological remission. This was consistent with the reports of the BFM Study Group [6,7]. Despite the improvement in the treatment outcome in periods I-IV, the percentage of relapses remained unsatisfactory high. A significant reduction of relapses was observed in the last period (AML-BFM 2012 Registry); however, the follow-up was much shorter (median 37 months) than in previous periods.
In addition to therapeutic protocol improvement, supportive care developed, including antifungal prophylaxis, broad spectrum antibiotic for neutropenic fever, and experience in management with emergencies like hyperleukocytosis or tumor lysis syndrome. All of this contributed to a reduction in early deaths and deaths in remission, and finally to an increase in overall survival. The significant role of supportive care in the improvement of the AML treatment outcome was reported by other authors [2,[8][9][10][11].
Identification of the prognostic factors enabled stratification to the risk groups. At the beginning, it was based on the treatment response and FAB classification. Further studies of the AML-BFM Group, as well as clinical trials worldwide, revealed the prognostic significance of numerous genetic abnormalities [12][13][14][15][16][17][18]. Since introduction of the protocol AML-BFM 2004, genetic abnormalities were taken into account in the stratification to risk groups (t(8;21), inv(16) as a favorable and FLT3-ITD as adverse). In the protocol AML-BFM 2012, the genetic abnormalities became crucial as the prognostic factors (Table 3). In our study, an analysis of survival within risk groups was performed for the last two periods. In the group treated with AML-BFM 2004 Interim, the probability of five-year OS was significantly higher in the SRG compared to HRG. Difference in the probability of EFS was almost significant between SRG and HRG, while no significant differences were found in the probability of RFS. Similarly, in the protocol AML-BFM 2012 Registry there was a significant difference in the OS and EFS between SRG and HRG, while no significant difference between the risk groups was noted in terms of RFS. According to the risk group criteria, all non-responders were classified to the HRG, which could mainly affect lower EFS or OS in that group, while the HR patients who managed to achieve morphological remission had a similar risk of relapse as the SR patients. The relatively low rate of relapses in the HRG could be the effect of HSCT recommended for the high-risk patients.
In our study, the results of the FLT3 analysis were available for 407 patients treated with AML-BFM 2004 and AML-BFM 2012, 41 of which (10%) were FLT3-ITD positive. That is similar to what has been reported by other authors [16,17]. In the Children's Cancer Group (CCG) study, FLT3-ITD was present in 77 of the 360 tested patients (12%) [16], and in the study performed by the AML-BFM Study Group, this mutation was found in 52 out of the 353 analyzed patients (15%) [17]. In our cohort, the probabilities of OS, EFS, and RFS were significantly lower in the FLT3-ITD positive children compared with the FLT3-ITD negative patients. There were no significant differences in OS, EFS, and RFS in the FLT3-ITD positive patients depending on the treatment protocol (AML-BFM 2004 vs. AML-BFM 2012). The treatment results in the analyzed FLT3-ITD group were similar to the results reported by other authors. In the study published by Nicoreth et al. in the group of 52 FLT3-ITD positive patients treated in Germany with AML-BFM 2004 and AML-BFM 2012, the probability of three-year OS and EFS survival was 0.54 ± 0.07 and 0.39 ± 0.07, respectively [17]. In the CCG study, FLT3-ITD positive patients treated on CCG-2941 and -2961 between September 1995 and December 2001 had four-year OS of 0.33 ± 0.13 [16]. Targeted treatment with FLT3 inhibitors has been studied in the last two decades with promising results [1,21,22]. Midostaurin, one of the first-generation FLT3 inhibitors, is the first tyrosine kinase inhibitor approved by the FDA for AML therapy in the first line, and is now being investigated in combination with chemotherapy in children with newly diagnosed FLT3-mutated AML [23].
Inv(16)(p13q22) is associated with acute myeloid leukemia subtype M4Eo and is known as a favorable prognostic factor. It results in the fusion of CBFB and MYH11 genes. The rearrangement is described in about 8% of pediatric AML [13,24]. In our study, there were 24 patients (6% of children with available results) with the CBFB-MYH11 fusion gene (12 in period IV and 12 in period V). No resistant disease was noted in that group of patients. In period IV, there were four relapses and three patients died (all because of toxicities). In period V, there were no relapses, and one patient died in remission because of severe infection. Probabilities of five-year OS, EFS, and RFS were 0.82 ± 0.08, 0.65 ± 0.11 and 0.74 ± 0.11, respectively. The outcome in patients with inv(16) treated with AML-BFM 2012 seemed to be better compared with the protocol AML-BFM 2004; however, the differences were not statistically significant, probably because of the low number of patients in the subgroups. The results are similar to those reported by other authors [12,13].
Translocation (8;21) with fusion gene RUNX1-RUNX1T1 is well known as a favorable prognostic factor in AML, and is detected in 12-14% of pediatric AML [13,25]. In the protocol, AML-BFM 2004 patients with t(8;21) were classified to SRG, and were treated without toxicities and from the disease progression was 10% and 13%, respectively. The results are similar to those described by other authors [32,33].

Conclusions
The use of standardized, systematically modified therapeutic protocols, with successive consideration of genetic prognostic factors in stratification into risk groups, and advances in supportive care significantly improved AML treatment outcomes over the past 40 years. Currently, the treatment results in pediatric AML in Poland are comparable to the results described by other research groups. Surprisingly good results were obtained in the group of patients with genetic abnormalities known as adverse prognostic factors like KMT2A -MLLT10/t(10;11)(p12;q23) and DEK-NUP214/t(6;9)(p23;q24), probably due to appropriate classification to the HRG with HSCT in the first CR. However, an unsatisfactory outcome was still found in the patients with FLT3-ITD. This group of patients may benefit in the future from targeted therapy with FLT3 inhibitors that are currently widely being investigated.
Continued close international cooperation, further improvement of the stratification into risk groups and indications for HSCT based on genetics and MRD assessment, and new drugs (especially targeted treatment) seem to be crucial directions for the future.