Real-World Validation of the European Leukemia Network 2022 Risk Stratification in Acute Myelogenous Leukemia

Background: The 2017 European LeukemiaNet (ELN 2017) guidelines have become foundational for the risk stratification, prognostication, and treatment of acute myelogenous leukemia (AML). The updated 2022 European LeukemiaNet (ELN 2022) guidelines include additional cytogenetic abnormalities and genetic mutations for risk group assignment. It remains unknown how these updated guidelines compare for risk stratification and prognosis. Methods: We studied (N=287) adults with newly diagnosed AML from MedStar Georgetown University, MedStar Washington Hospital Center, and Baylor University Medical Center from 2012 to June 2023. Patients were stratified as favorable (fav), intermediate (int), or adverse (adv) risk according to ELN 2017 and ELN 2022 criteria. Overall survival (OS) was compared by age, race, and sex. Statistical analysis performed with SAS. Results: Median age was 65­+16 years with a predominance of males (N=168, 59%). African-Americans (AA) composed 15% (N=44) of the cohort. Myelodysplastic-related mutations were present in 33 patients (15%), whereas the prevalence of epigenetic mutations was 25% (N=55). By ELN 2017 criteria, 19 (7%) patients were characterized as fav, 64 (23%) as int, and 198 (70%) as adv risk. Using ELN 2022 criteria, 18 (6%) patients were categorized as fav, 52 (19%) as int, and 211 (75%) as adv risk. Kaplan-Meier (KM) curves revealed a median overall survival of 200 days (d), 400 d, and 1532 d in the adv, int, and fav risk groups, respectively (log-rank p<0.001) by ELN 2017 criteria. Median OS was 379 d, 527 d, and 1552 d for the adv, int, and fav groups, respectively (log-rank p<0.001) in the ELN 2022 cohort. Stratifying by age <60 years, median OS was 390 d in the adv group and 832 d in the int group (log-rank p=.021) by ELN 2017 criteria, as compared to 350 d in the adv group and 757 d in the int group (log-rank p=.0012) by ELN 2022 criteria. Among those >60 years of age, median OS was 244 d and 349 d in the adv and int groups, respectively by ELN 2017 criteria (Anova p<0.25). In the ELN 2022 cohort, median OS among those >60 years of age was 284 d for the adv group and 209 days in the int group (Anova p<0.62). Among females, median OS in the adv group was 224 d and 342 d in the int group by ELN 2017 criteria (Anova p=0.4). In the ELN 2022 cohort, median OS among females was 214 d in the adv group and 421 d in the int group (Anova p=.13). Median OS among male patients in the ELN 2017 cohort was 324 d for those of adv risk and 623 d for those of int risk (Anova p<0.004). In the ELN 2022 cohort, median OS in males was 352 d and 563 d in the adv and int groups, respectively (Anova p<0.07). In the ELN 2017 cohort, median OS was 259 d in AA patients as compared to 719 d in non-AA patients; this was not statistically significant (Anova p=0.16). Comparatively, in the ELN 2022 cohort, median OS for the AA patients was 153 d as compared to 212 d for the non-AA patient population (Anova p=0.86). Conclusions: In this real-world dataset, the updated ELN 2022 guidelines reproduce differential survival for proposed favorable, intermediate, and adverse risk categories. When considering time to endpoint survival [Kaplan Meier curves] in our study, ELN 2022 seems to lack the ability to clearly discriminate among the adverse and intermediate risk subgroups. It is possible that patients seen in academic institutions not participating in FLT3 directed therapy trials retained inferior survival whereas similar patients fared better through early enrollment in trials using molecular targeted therapy. Finally, neither ELN 2017 nor ELN 2022 discriminate survival in AA populations for either adverse or intermediate subgroups. Efforts should be directed to develop novel risk stratification schemas to better encapsulate this vulnerable population.