Advancements in understanding acute myeloid leukemia (AML) genetics have led to new diagnostic entities and improved prognostic system [1,2,3,4]. The European LeukemiaNet (ELN) group updated prognostic stratification in 2022, which has been validated in several chemotherapeutic AML cohorts [5,6,7,8]. However, the applicability of ELN-2022 risk system in AML patients undergoing allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains uncertain. Our study aims to shed light on this.

We reclassified 600 AML patients who underwent allo-HSCT by ELN-2022 genetic risk categories: 214 (35.67%) were favorable-risk, 162 (27.0%) were intermediate-risk and 224 (37.33%) were adverse-risk. Eighty-six (14.33%) patients shifted from ELN-2017 risk stratification (Fig. 1A, B). Reasons for these shifts are detailed in Additional file 1: Table S1.

Patients and genetic characteristics and impact of ELN-2022 and ELN-2017 risk stratification on clinical outcomes. A Relationship of risk groups between ELN-2022 and ELN-2017 risk groups; B Distribution of re-stratification in ELN-2017 risk groups. C Landscape of genetic abnormalities defined by ELN-2022 genetic risk categories. The color scale is representative of a number of patients. D Additional mutations stratified by ELN-2022 genetic risk categories. Genes mutated in more than ten patients are shown. E Overall survival stratified by ELN-2022 risk categories. F Overall survival stratified by ELN-2017 risk categories. G Cumulative incidence of relapse stratified by ELN-2022 risk categories. H Cumulative incidence of relapse stratified by ELN-2017 risk categories

We assessed the frequency of genetic abnormalities defined by ELN-2022 and the distribution of additional genes mutated in more than 10 patients (Fig. 1C, D). Correlation analysis showed that t(8;21) strongly correlated with KIT mutation (r = 0.5, P < 0.001), SF3B1 mutation strongly correlated with inv(3) (r = 0.5, P < 0.001).

Patients and transplant-related characteristics were listed in Additional file 1: Table S2. Compared to favorable- and intermediate-risk groups, adverse-risk group had a lower percentage of bone-marrow blasts at initial diagnosis (P = 0.036) and a higher proportion of refractory/relapse- and secondary-AML (P = 0.006, p < 0.001, respectively, Additional file 1: Fig. S1).

The three-year and five-year overall survival (OS), event-free survival, cumulative incidence of relapse (CIR) and non-relapse mortality stratified by ELN-2022 and ELN-2017 are shown in Additional file 1: Table S3. Compared to favorable-risk, OS shortened significantly as the ELN-2022 risk stratification increased but didn’t significantly in ELN-2017 intermediate-risk (Fig. 1E, F). Pairwise comparisons for OS revealed significant differences between the ELN-2022 favorable- and intermediate-risk groups (P = 0.047) but not between the intermediate- and adverse-risk groups (P = 0.455). Based on ELN-2017 risk stratification, OS was not significantly different between intermediate- and favorable-risk groups (P = 0.115) or between intermediate- and adverse-risk groups (P = 0.115). Both ELN-2022 and ELN-2017 adverse-risk were associated with increased CIR. (Fig. 1G, H) Smoothed hazard estimates showed a higher mortality risk within 6 months post-transplantation in ELN-2022 intermediate-risk group than in adverse-risk group. Assessment based on ELN-2017 recommendations indicated that adverse-risk group had the highest hazard ratio for death in 1-year post-transplantation, followed by intermediate- and favorable-risk groups (Additional file 1: Fig. S2).

We performed time-dependent receiver operating characteristic (ROC) analysis to validate the prognostic efficacy of ELN-2022 and ELN-2017 risk systems in our transplant cohort. The AUC for predicting OS gradually increased from one to five years post-transplantation, with the AUC for ELN-2022 consistently higher than of ELN-2017 (Fig. 2A). However, AUC for 3-year and 5-year OS between two ELN versions was not significantly different (P = 0.458, P = 0.838, respectively).

Impact of MRD pre-transplantation and MRD-modified ELN-2022 risk system on clinical outcomes. A Dynastic AUC for overall survival at different time points after transplantation according to ELN-2017, ELN-2022 and MRD-modified ELN-2022 risk system. B Impact of MRD pre-transplantation on overall survival. C Impact of MRD pre-transplantation on cumulative incidence of relapse. D Impact of MRD pre-transplantation combined with ELN-2022 risk stratification on overall survival. E Impact of MRD pre-transplantation combined with ELN-2022 risk stratification on cumulative incidence of relapse. F Relationship of risk groups between ELN-2022 and MRD-modified ELN-2022 risk groups. G Overall survival according to MRD-modified ELN-2022 risk groups. H Cumulative incidence of relapse according to MRD-modified ELN-2022 risk system

We separated patients into three groups based on pre-transplant minimal residual disease (MRD): MRD-negative (395, 65.8%), MRD-positive (90, 15.0%) and not-CR (115, 19.2%). Median survival was not reached for MRD-negative, 3.70 (95% CI 1.6 to NA) years for MRD-positive and 2.07 (95% CI 1.35 to 4.73) years for not-CR patients (P < 0.001) (Fig. 2B, C). Further stratification based on both MRD and ELN-2022 was conducted. The survival of MRD-negative patients in favorable- and intermediate-risk groups was comparable and longer than adverse-risk group. OS and CIR of MRD-positive patients were not significantly different among the three ELN-2022 groups and were similar to Not-CR patients (Fig. 2D, E). Based on aforementioned analysis, we created the MRD-modified ELN-2022 risk system for transplant AML patients. Number and risk-shift of patients from ELN-2022 risk groups to MRD-modified risk groups are shown in Fig. 2F. Three-year OS after transplantation of modified low-, intermediate- and high-risk was 79.5% (95% CI 74.4% to 84.9%), 63.69% (95% CI 55.01% to 73.74%), 47.77% (95% CI 40.79% to 55.94%) (P < 0.001) and 3-year CIR after transplantation was 10.09% (95% CI 6.53% to 14.55%), 23.76 (95% CI 15.82% to 31.79%), 40.65 (95% CI 33.28% to 47.88%) (P < 0.001, Fig. 2G, H). Time-dependent ROC analysis for 3-year survival significantly outperforms ELN-2022 (68.23% vs 53.31%, P < 0.001), as well as for 5-year survival (72.81% vs 58.80%, P < 0.001, Fig. 2A).

In conclusion, ELN-2022 risk system had superior separation for survival of favorable- and unfavorable-risk groups but poor separate for intermediate- and adverse-risk groups. ELN-2017 risk system primarily separates survival of favorable- and adverse-risk groups. Both ELN-2022 and ELN-2017 systems exhibited limited prognostic utility for AML patients undergoing allo-HSCT. Pre-transplant MRD provides additional prognostic insights and MRD-modified ELN-2022 risk system enhances prognostic ability for transplantation.