Introduction: Racial/ethnic disparities in acute myeloid leukemia (AML) have been increasingly recognized. However, data outside of Black or Hispanic populations is extremely limited. Recent research efforts have dramatically improved our ability to risk stratify and manage AML. Although with the long-standing over-representation of White patients on clinical trials (Hantel A, et al. Blood Adv. 2021;5(21):4352-4360), which our current guidelines are based on, the generalizability of such guidelines to other racial/ethnic groups is unknown. Better insight on the disparities in AML from both biological and non-biological aspects is crucially needed in order to identify actionable targets for intervention. In this study, we aim to examine the previously unknown outcome disparity between White and Asian patients with newly diagnosed AML (ND-AML) and to explore the potential impact of differences in AML genomics.

Methods: Adult patients with ND-AML cared at Memorial Sloan Kettering Cancer Center who self-identified as White or Asian and had comprehensive genomic workup at diagnosis including MSK-IMPACT HEME, a targeted disease-focused sequencing test (Ptashkin RN, et al. Nat Commun. 2023;14(1)6895), were included. Overall survival (OS) was defined from the time of diagnosis to the time of death of any cause. Kaplan-Meier method and Cox regression modeling were used to compare the survival function and hazard of death between Asian and White patients.

Results: A total of 437 patients with AML diagnosed between 2017 to 2024 were included (53 Asian, 384 White). Asian patients were younger at diagnosis compared to Whites (median age 58 vs. 66 years) with similar distribution of AML subtypes (de novo AML 57% vs. 54%) and prevalence of adverse risk disease defined by European LeukemiaNet (ELN) 2017 criteria (53% vs. 48%). At median follow-up of 41 months, Asians had significantly longer OS relative to Whites (median OS (mOS) 64.6 vs. 21.5 months, p=0.01). The OS disparity observed was partially confounded by age and AML subtype and attenuated over time, suggesting a greater impact on early mortality (adjusted HR 0.40, 95% CI 0.19-0.84 at time 0 and increased by 1.03 for every month increase in time. Ref: White). The OS disparity observed was maintained after accounting allogeneic stem cell transplant as a time-varying covariate; and when stratified by induction intensity.

ELN2017 criteria successfully risk stratified White patients (mOS: unreached vs. 42.6 vs. 12.9 months in favorable, intermediate and adverse group, respectively; p<0.001). However, it did not differentiate intermediate and adverse risk groups in Asian patients (mOS: unreached vs. 32.2 vs. 38.8 months, respectively; p=0.14). Similar results were observed when using ELN2022 criteria. Between racial groups, the largest OS difference was observed in patients in ELN2017 adverse group with mOS being 38.8 months in Asians and 12.9 months in Whites, respectively (age and AML type adjusted HR: 0.51, 95% CI 0.29-0.90. Ref: White). Within the ELN2017 adverse group, Asian patients had numerically more adverse cytogenetics (75% vs. 67%), but less TP53 alterations (25% vs. 37%; bi-allelic TP53 14% vs. 26%). Asian patients with ELN2017 adverse risk disease without TP53 alterations had particularly favorable outcomes (mOS 44.9 vs. 17.4 months, p=0.01), whereas those with TP53 alterations had similar OS relative to Whites (mOS 8.3 vs. 7.9 months, p=0.45).

Conclusions: In this comprehensive, real-world cohort, we demonstrate for the first time outcome disparities between Asian and White patients with ND AML. Similar to observations in Blacks and Hispanics, Asian patients are diagnosed with AML at a younger age compared to Whites. Asian patients have superior OS relative to Whites in this cohort. Intriguingly, ELN prognostic classifiers did not adequately stratify OS outcomes for Asian patients. Outside of TP53 alterations, Asian patients classified as “adverse risk” appear to have quite favorable OS than historical data from mostly White patients. Ongoing analyses are focused on examining the relapse-related outcomes and details in the baseline genomic landscape to have a better understanding of the mechanisms of the OS disparity observed. Our results suggest a possible need for race/ancestry-inclusive considerations for AML risk stratification and management strategies, which may expand the current scope of AML disparity research.

* XW and YL are co-senior authors.

This content is only available as a PDF.
2025
Sign in via your Institution