Molecular risk stratification and MRD-guided consolidation determine outcomes in pediatric non-down syndrome acute megakaryoblastic leukemia: A multicentre Chinese cohort study

Abstract Background: Management of pediatric acute megakaryoblastic leukemia (AMKL) without Down syndrome (non-DS-AMKL) remains challenging due to suboptimal induction responses and poor survival. This study assessed the efficacy of fludarabine, cytarabine, granulocyte colony-stimulating factor, and idarubicin (FLAG-IDA) in non-DS-AMKL and to evaluate the role of risk-adapted consolidation strategies, including complete remission (CR1) hematopoietic stem cell transplantation (HSCT), guided by genetic profiles and measurable residual disease (MRD) dynamics. Methods: In this multicentre retrospective study from China, we analyzed 58 non-DS-AMKL patients treated with FLAG-IDA (n = 50) or daunorubicin, cytarabine, etoposide (DAE) (n = 8) induction, followed by risk-adapted consolidation. High risk (HR): defined by CBFA2T3::GLIS2, NUP98::KDM5A, KMT2A rearrangements (KMT2A-r), FUS::ERG, monosomy 7 (-7), or FLT3-ITD. Non-HR: all remaining cases that did not meet the HR criteria. This classification schema was prospectively applied to analyze associations between risk-adapted consolidation therapies (chemotherapy alone vs. HSCT) and survival outcomes. Results: (1)Patient characteristics: In the C-HUANAN-AML cohort of 875 patients with pediatric de novo AML (excluding acute promyelocytic leukemia), 58 (6.6%) had non-DS-AMKL. Baseline characteristics compared with other AML subtypes revealed patients with non-DS-AMKL were significantly younger (median 23 vs 79 months; p<0.001), more frequently female (55.2% vs 41.5%; p=0.043), and had lower median WBC count (11.5 × 10⁹/L vs 22.2 × 10⁹/L; p<0.001). (2)Overall long-term outcomes: With a median follow-up of 39.9 months (range, 1–103.4 months), the 5-year OS and EFS of the patients with non-DS-AMKL were significantly lower than those of children with other AML subtypes (OS: 61.7% ± 6.4% vs. 78.2%,p = 0.001; EFS: 58.5% vs. 68.8%, p = 0.045) due to higher relapse (34.8% vs. 19.5%, p = 0.002). Compared to the patients who received DAE induction (Group B), the patients treated with FLAG-IDA induction did not have superior outcomes. (3)Prognostic impact of clinical and genetic factors: To mitigate potential statistical power limitations arising from small subgroup sizes, we aggregated molecular subtypes with analogous prognoses into risk-stratified cohorts as predefined (Methods). The HR group demonstrated significantly inferior 5-year OS (42.1% vs 75.8%; p = 0.023) and EFS (EFS: 39.0% vs 72.7%; p = 0.023) compared to the non-HR group, coupled with a 2.8-fold increased CIR (55.6% vs 20.0%; p = 0.011). (4)Prognostic impact of the induction chemotherapy response​​: Morphologic response (achievement of CR) could significantly stratify OS and EFS at both the first-course (time point 1, TP1) and second course (time point 2, TP2). MFC-MRD assessments (<0.1%) at TP1 could not significantly stratify OS and EFS, whereas MFC-MRD at TP2 independently stratified both OS and EFS (OS: 73.5% vs. 25.0%, p < 0.001; EFS: 69.2% vs. 25.0%, p < 0.001). Notably, MRD status at TP2 emerged as a critical discriminator: those achieving CR with MRD-negative status demonstrated significantly superior OS and EFS compared to CR patients with MRD-positive status at TP2 (OS: 74.7% vs. 25.0%, p = 0.007; EFS: 70.2% vs. 25.0%, p = 0.019). (5)Prognostic impact of post-induction treatment: In the overall cohort, HSCT during CR1 significantly improved outcomes compared to consolidation chemotherapy alone (OS: 84.8% vs. 53.8%,p = 0.012; EFS: 84.8%vs. 45.8%, p = 0.002). In the predefined HR genetic subgroup, HSCT in CR1 conferred significant survival advantages over chemotherapy alone: 5-year OS (74.1% vs. 16.7%,p = 0.006) and EFS (74.1%% vs. 11.1%,p < 0.001); in contrast, this benefit was not observed in the non-HR genetic subgroup. Notably, patients who failed to achieve morphologic CR and/or had persistent MRD positivity at TP1 demonstrated significantly improved 5-year OS (71.6% vs. 33.3%,p = 0.034) and EFS (71.6% vs. 16.7%,p = 0.004) when undergoing HSCT during CR1 compared to chemotherapy alone as consolidation. Conclusion: While induction intensification with FLAG-IDA offers no survival advantage over DAE, risk-stratified HSCT and MRD-guided consolidation are indispensable for mitigating relapse. Future protocols must prioritize molecular characterization at diagnosis and accelerate the translation of novel agents targeting disease-specific vulnerabilities.