“As Treated” Validation of the European LeukemiaNet (ELN) Classification for Adult Acute Myeloid Leukemia (AML) Patients (pts) Treated in a Large AML-Referral Center from 2010-2022

Introduction: The ELN classification of AML incorporates cytogenetic and mutational risk profiles of AML to define separate favorable, intermediate and adverse risk groups. Recommendations for post remission treatment (PRT) intensity, most importantly allogeneic stem cell transplantation (alloSCT), are based on these risk groups. The ELN classification is regularly updated as the understanding of genetic risk factors for AML prognosis improves. In-depth genetically defined cohorts from registries or clinical trials, often reported as retrospective real-world validation cohorts, contain an inherent bias as pts were not treated according to treatment recommendations corresponding to the time interval of the respective ELN classification.

Methods: 662 AML pts who received intensive induction therapy at the University Hospital Muenster between 2010 and 2022 were included in this retrospective analysis. Pts were generally classified according to ELN 2010 recommendations if treated between 2010-2016 (ELN10), and ELN 2017 recommendations if treated between 2017-2022 (ELN17). Survival was estimated according to Kaplan-Meier. Cumulative incidences of relapse (CIR) and non-relapse mortality (NRM) were calculated using Gray’s method. To eliminate treatment bias, we retrospectively reclassified the risk of the ELN17 group according to ELN 2010 (ELN10_re) and compared outcomes with the actual ELN10 cohort in identical risk groups.

Results: Baseline characteristics of pts were evenly distributed between ELN10 (n=373) and ELN17 (n=289), except for higher proportions of therapy-related and intermediate risk AML after 2016. Median follow-up was 61 months, with 90 months for ELN10 and 42 months for ELN17 cohorts, respectively. Overall survival (OS), event-free survival (EFS) and relapse-free survival (RFS) significantly improved for ELN17 pts compared to ELN10 (4-year OS 60% vs. 40%, HR 0.59, 95% CI 0.5-0.7, p<.0001; 4-year EFS 37% vs. 21%, HR 0.68, 95% CI 0.6-0.8, p<.0001; 4-year RFS 54% vs. 35%, HR 0.61, 95% CI 0.5-0.8, p=.00026). ELN17 pts achieving a first complete remission (CR1) had a significantly reduced CIR with similar NRM compared to ELN10 (CIR, HR 0.57, CI 0.4-0.8, p<.001; NRM, HR 0.95, CI 0.6-1.6, p=.86). Generally, alloSCT as PRT was recommended for pts with ELN10 intermediate risk (HLA match-related donor) and adverse risk (HLA match-related or unrelated donor) AML. With the implementation of the ELN 2017 classification, alloSCT as PRT was recommended for pts with intermediate and adverse risk irrespective of donor relation. Comparing time intervals, ELN17 pts receiving an alloSCT either in CR1 (n=211), in primary refractory situation after at least one induction cycle (n=124) or following relapse (n=100), showed better OS and RFS (4-year OS_allo 66% vs. 50%, HR 0.62, 95% CI 0.5-0.8, p=.0016; 4-year RFS_allo 59% vs. 45%, HR 0.68, 95% CI 0.5-0.9, p=.0056) and lower NRM rates (HR 0.63, CI 0.4-0.9, p=.02), while CIR was unchanged (HR 0.78, CI 0.5-1.1, p=.2). To eliminate treatment bias, retrospectively reclassified ELN10_re pts of the ELN17 group were compared to regularly classified ELN10 pts. Risk distribution according to ELN 2010 criteria was comparable between time intervals. The majority of ELN17 pts were retrospectively reclassified from favorable (23 pts /8% of cohort) or adverse (58 pts / 20%) to ELN10_re intermediate risk, while no patient was reclassified to ELN10_re adverse risk. ELN10_re intermediate risk pts, treated between 2017-2022 benefited most in OS, EFS and RFS (OS, HR 0.5, 95% CI 0.4-0.7, p<.0001; EFS, HR 0.58, 95% CI 0.5-0.8, p<.0001; RFS, HR 0.47, 95% CI 0.3-0.7, p<.0001). ELN10_re favorable and adverse group comparisons showed no significant difference in outcome over time intervals.

Conclusion: Based on the analysis of more than 600 AML pts treated at our center from 2010-2022, we conclude that since 2017, besides approval of novel substances for conventional induction and consolidation therapies, the update of the ELN classification - moving more pts to intermediate and adverse risk groups with subsequent intensification of PRT through a higher proportion of pts proceeding to alloSCT – significantly improved the survival of AML pts. This improvement is particularly pronounced for intermediate risk pts according to ELN 2010 criteria. To the best of our knowledge, this is the first “as treated” real-world validation for the ELN 2017 update reported.