Treatment-specific prognostic performance of ELN 2022, ELN 2024, and Beat AML 2024: a real-world AML validation study.

Real-world validation of the European LeukemiaNet (ELN) 2022, ELN 2024, and Beat Acute Myeloid Leukemia (AML) 2024 across different therapies is needed. We evaluated 426 patients with AML (median age, 68 years; 60% male) treated from 2016 to 2025. ELN 2024 (55%, 24%, and 20%), ELN 2022 (20%, 20%, and 60%), and Beat AML 2024 (21%, 38%, and 41%) assigned patients to favorable-, intermediate-, and adverse-risk groups with variation. Less-intensive treatment (LIT) had lower complete remission (CR) odds than intensive chemotherapy (IC). For CR prediction, ELN 2022 (area under the receiver operating characteristic curve [AUC], 0.763) and Beat AML 2024 (AUC, 0.759) outperformed ELN 2024 (AUC, 0.718). With a median follow-up of 49.8 months, the median overall survival (mOS) was 15.7 months. Each system stratified mOS (all P < .001); overall c-index favored ELN 2022 (0.640) and Beat AML 2024 (0.635) over ELN 2024 (0.601). For patients treated with IC, ELN 2022 (0.680) and Beat AML 2024 (0.664) exceeded ELN 2024 (0.601); for patients treated with LIT, ELN 2024 and Beat AML 2024 identified a favorable group with better mOS, whereas ELN 2022 failed to discriminate between favorable and intermediate risk. Allogeneic hematopoietic stem cell transplantation improved mOS; in time-dependent analysis, adverse risk by Beat AML 2024 (hazard ratio [HR], 4.0), ELN 2024 (HR, 3.0), and ELN 2022 (HR, 2.5) remained independently unfavorable. Overall, Beat AML 2024 was the most informative in IC, and ELN 2024/Beat AML 2024 better identified favorable patients in LIT.