Abstract
Acute myeloid leukemia (AML) is a biologically heterogeneous hematologic malignancy, and racial disparities in outcomes remain incompletely characterized. This retrospective study analyzed 677 patients diagnosed with AML at Moffitt Cancer Center (2014–2024) to evaluate the impact of race on genomic profiles and clinical outcomes. The cohort comprised 18.6% Black, 19.1% Hispanic, and 62.3% White patients. While ELN 2022 risk groups and most driver mutations were similarly distributed, Black patients exhibited a higher frequency of KRAS mutations (p < 0.05). With a median follow-up of 3.3 years, median overall survival (OS) was 2.2 years for White, 1.5 years for Black, and 2.4 years for Hispanic patients (p = 0.045). Notably, no survival differences existed within favorable or adverse ELN risk categories; however, Black patients with intermediate-risk disease experienced significantly inferior OS compared to White (p = 0.02) and Hispanic patients (p = 0.003). Among those receiving intensive induction, Black patients had significantly worse OS (p ≤ 0.001), while Hispanic patients treated with lower-intensity regimens demonstrated inferior OS compared to White patients (p = 0.01). In multivariable analysis, race was not an independent predictor of OS. Survival was independently driven by prior myeloid malignancy, ELN risk, allogeneic HCT, and mutations in TP53 and IDH2. Regarding race-specific molecular drivers, TP53 and IDH2 mutations were associated with OS in White patients, TP53 in Black patients, and IDH1 in Hispanic patients. These findings suggest that while racial disparities exist in specific clinical subgroups, they are primarily driven by distinct molecular profiles, cytogenetic heterogeneity, and patient-specific clinical factors rather than race alone.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s00277-026-07145-3.
Keywords: Acute myeloid leukemia, Race, Overall survival, Mutation, Treatment intensity, ELN risk
Introduction
Acute myeloid leukemia (AML) is a clonal disorder marked by impaired differentiation, leukemic blast expansion, and suppressed normal hematopoiesis [1]. According to the Surveillance, Epidemiology, and End Results (SEER) database, an estimated 22,010 new AML cases were expected in the United States in 2025, representing roughly 1.1% of all new cancer diagnoses [2]. Although targeted therapies and newer combination strategies have yielded incremental survival gains, illustrated by a rise in 5-year overall survival (OS) from 22% in 2004–2010 to 28% in 2011–2019 in a National Cancer Database (NCDB) analysis, AML remains a challenging disease to manage [3]. High early mortality, treatment-related complications, and difficulty achieving durable remission continue to limit better outcomes [4, 5].
Multiple factors influence survival in AML, including age, comorbidities, socioeconomic status, and molecular alterations [6–11]. Race has also been associated with differences in disease biology and outcomes. Several studies have suggested that Black patients may have distinct driver mutations and inferior OS compared to White patients [12–14]. Hispanic patients, who represent another significant US minority, have also been reported in some analyses to experience worse outcomes than non-Hispanics [15]. However, other studies have shown mixed or conflicting findings, with no consistent survival differences among racial groups [5–17]. The mutational landscape of AML across racial groups and the influence of demographic, molecular, and treatment factors on outcomes warrant further investigation.
This study describes outcomes in an institutional cohort of adult AML patients treated between 2014 and 2024. Specifically, we aim to (1) characterize the mutational landscape by race, and (2) report median and 5-year OS while examining demographic, disease-related, and treatment variables associated with survival after AML diagnosis across racial groups.
Methods
Patient selection and data source
We conducted a retrospective review of data from 677 adults with AML evaluated at Moffitt Cancer Center (MCC). The MCC catchment area comprises 23 counties across West Central Florida and encompasses approximately 47% of the state’s population. The catchment area includes all counties within two Florida Cancer Control Collaborative regions: East Central Florida and Southwest Florida. Compared with the overall populations of Florida and the United States (US), the catchment area has a lower proportion of Black residents. The proportion of Hispanic residents is lower than that of Florida overall but higher than that of the US overall. Additionally, the proportion of individuals living below the poverty level is comparable to both Florida and national estimates. With respect to referral patterns, most patients are referred from community oncology centers located both within and outside the MCC catchment area throughout Florida, as well as from neighboring states [18].
All patients aged ≥ 18 years with a pathologically confirmed AML diagnosis are included in the institutional AML database. Risk stratification followed the 2022 European LeukemiaNet (ELN) risk classification. For Black and Hispanic patients, individuals diagnosed between 2014 and 2024 were included. For White patients, those diagnosed between 2020 and 2024 were analyzed to maximize the number of patients with comprehensive molecular information. Patients with insufficient cytogenetic/mutational data for ELN 2022 classification were excluded. This study was approved by the Institutional Review Board of MCC.
Cytogenetic/mutational profiling
Conventional G-banded karyotyping on bone marrow aspirates analyzed 20 metaphases. We confirmed clonal abnormalities by the presence of at least two affected cells. Somatic mutations were identified using next-generation sequencing (NGS) performed on bone marrow or peripheral blood samples. Sequencing platforms included Illumina TruSight Myeloid 54-gene panel, Genoptix NexCourse Complete, Genoptix Myeloid Molecular Profile, FoundationOne Heme, and the Moffitt 98-gene Myeloid Action Panel that was used after 2020 (list of genes included in Supplemental Sect. 1).
Variables
We collected details on the following variables: age at diagnosis, sex, race, Eastern Cooperative Oncology Group (ECOG) performance status (PS), prior myeloid or other malignancies with treatment history, extramedullary disease, frontline treatment and intensity, hematopoietic cell transplantation (HCT), date of diagnostic bone marrow (BM) biopsy, aspirate cellularity and blast percentage, BM reticulin fibrosis, white blood cell count (WBC), absolute neutrophil count (ANC), hemoglobin, platelet count, circulating blast percentage, cytogenetics, fluorescence in situ hybridization (FISH) results, gene mutations, and vital status.
Statistical analysis
Baseline characteristics were summarized as counts and percentages for categorical variables and as means with standard deviations or medians with interquartile ranges for continuous variables. Group comparisons for categorical variables used a Chi-square test or Fisher’s exact test, with pairwise p-values adjusted for multiple comparisons using the Benjamini–Hochberg false discovery rate method. Continuous variables were compared using t-tests. To account for temporal bias arising from differing enrollment periods, a subgroup analysis was conducted exclusively among patients diagnosed between 2020 and 2024 to align with the inclusion timeframe of the White patient cohort.
OS was defined as the time from AML diagnosis to death from any cause, with survivors censored at last known follow-up. Median, 3-year, 5-year OS were calculated using Kaplan-Meier method from AML diagnosis to death, loss to follow-up, or the censor date (June 30, 2025). Differences in OS between groups were evaluated using the log-rank test. For univariate Cox proportional hazards models, gene mutations were included only if at least five events were available. Multivariable Cox models used complete cases only (any predictors with missing value were excluded) and employed backward selection based on the Akaike Information Criterion (AIC). Factors known to influence AML outcomes were included a priori regardless of statistical significance besides additional covariates from univariate analysis. All statistical analyses were conducted using R version 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria). P-value < 0.05 was considered statistically significant.
Results
Characteristics of patients included in analysis
We included 677 patients diagnosed with AML from 2014 to 2024. The median age at diagnosis was 65 (range: 16–93) years. There were more males than females (54.3% vs. 45.7%), and White was the most common race (62.3%). Approximately 24% of patients had prior myeloid malignancy. More than 92% of patients had ECOG PS of 0–1. Patients were risk stratified based on 2022 ELN classification, and the majority (54.7%) had adverse risk AML. Approximately 61% of patients were treated with an intensive regimen, and 35.8% eventually underwent allogeneic HCT (Table 1).
Table 1.
Baseline characteristics of the study population by race
| Variables | Full cohort (N = 677) |
White (N = 422) |
Black (N = 126) |
Hispanic (N = 129) |
p (White vs. Black) | p (White vs. Hispanic) | p (Black vs. Hispanic) |
|---|---|---|---|---|---|---|---|
| Gender | 0.019 | 0.106 | 0.524 | ||||
| Female | 309 (45.7%) | 174 (41.2%) | 70 (55.6%) | 65 (50.8%) | |||
| Male | 367 (54.3%) | 248 (58.8%) | 56 (44.4%) | 63 (49.2%) | |||
| Age at diagnosis (years, range) | 65 (16–93) | 68 (21–93) | 61 (18–92) | 55 (16–91) | < 0.001 | < 0.001 | 0.422 |
| ECOG | 0.122 | 0.259 | 0.770 | ||||
| 0–1 | 627 (92.8%) | 399 (94.5%) | 111 (88.8%) | 117 (90.7%) | |||
| ≥ 2 | 49 (7.2%) | 23 (5.5%) | 14 (11.2%) | 12 (9.3%) | |||
| Prior malignancy | 92 (13.7%) | 60 (14.4%) | 21 (16.7%) | 11 (8.5%) | 0.634 | 0.168 | 0.168 |
| Prior myeloid malignancy | 159 (23.5%) | 104 (24.6%) | 30 (23.8%) | 25 (19.4%) | 0.942 | 0.719 | 0.719 |
| Type of prior myeloid malignancy | 0.929 | 0.929 | 0.929 | ||||
| CMML | 17 (10.7%) | 10 (9.6%) | 5 (16.7%) | 2 (8%) | |||
| MDS | 101 (63.5%) | 65 (62.5%) | 20 (66.7%) | 16 (64%) | |||
| MDS/MPN | 8 (5.0%) | 7 (6.7%) | 1 (3.3%) | 0 | |||
| Primary myelofibrosis | 11 (6.9%) | 7 (6.7%) | 2 (6.7%) | 2 (8%) | |||
| PV/ET | 18 (11.3%) | 14 (13.5%) | 2 (6.7%) | 2 (8%) | |||
| Unknown | 4 (2.6%) | 1 (1%) | 0 | 3 (12%) | |||
| WBC (/mm3), median (range) | 7.6 (0.40–403) | 6.9 (0.40–403) | 6.7 (0.63–315) | 11.2 (0.73–338) | 0.972 | 0.010 | 0.035 |
| Hemoglobin (g/dL), median | 8.5 (4.4–18.3) | 8.5 (4.6–15.6) | 8.2 (4.4–14.9) | 8.35 (4.9–18.3) | 0.717 | 0.717 | 0.781 |
| Platelet (/mm3), median (range) | 50 (2–1502) | 56 (2–1502) | 42 (3–925) | 44 (6–703) | 0.048 | 0.048 | 0.683 |
| Circulating blast % (range) | 20 (0–98) | 20(0–98) | 20(0–96) | 25(0–96) | 0.562 | 0.156 | 0.428 |
| Bone marrow aspirate blast % (range) | 54 (0–98) | 50 (0–98) | 55 (0–96) | 62 (10–95) | 0.103 | 0.027 | 0.374 |
| Reticulin fibrosis | 0.374 | 0.031 | 0.046 | ||||
| No fibrosis | 387 (63.8%) | 245 (61.6%) | 70 (61.4%) | 72 (75.8%) | |||
| Grade 1 | 162 (26.7%) | 118 (29.6%) | 29 (25.4%) | 15 (15.8%) | |||
| Grade 2 | 39 (6.43%) | 24 (6.03%) | 12 (10.5%) | 3 (3.16%) | |||
| Grade 3 | 19 (3.13%) | 11 (2.76%) | 3 (2.63%) | 5 (5.26%) | |||
| Extramedullary disease | 49 (7.3%) | 27 (6.5%) | 11 (9.2%) | 11 (8.5%) | 0.791 | 0.791 | 1.000 |
| ELN risk | 0.225 | 0.511 | 0.219 | ||||
| Favorable | 143 (21.2%) | 94 (22.3%) | 19 (15.1%) | 30 (23.3%) | |||
| Intermediate | 163 (24.1%) | 99 (23.5%) | 28 (22.2%) | 36 (27.9%) | |||
| Adverse | 370 (54.7%) | 228 (54.2%) | 79 (62.7%) | 63 (48.8%) | |||
| Frontline treatment intensity | 0.007 | < 0.001 | 0.212 | ||||
| Intensive | 412 (60.9%) | 228 (54%) | 87 (69%) | 97 (75.2%) | |||
| Low intensity (including HMA/venetoclax) | 225 (33.2%) | 168 (39.8%) | 33 (26.2%) | 24 (18.6%) | |||
| Supportive care/Unknown | 40 (5.9%) | 26 (6.2%) | 6 (4.8%) | 8 (6.2%) | |||
| HMA/venetoclax | 154 (22.7%) | 123 (29.2%) | 20 (15.9%) | 11 (8.5%) | 0.006 | < 0.001 | 0.109 |
| Allogeneic HCT at CR1 | 241 (35.8%) | 154 (36.7%) | 37 (29.8%) | 50 (38.8%) | 0.294 | 0.602 | 0.294 |
Patients of non-White race were significantly younger: White patient mean age at diagnosis, 68 years vs. Black 61 and vs. Hispanic 55 (p < 0.001). Approximately 25% of White patients, 24% Black, and 19% Hispanic had an antecedent myeloid malignancy (p = 0.47). A higher BM blast percent was noted in Hispanic patients compared to White (62% vs. 50%, p = 0.027). A lower proportion of Hispanics had grade ≥ 1 reticulin fibrosis compared to White and Black patients: Hispanic 24.2% vs. White 38.4% vs. Black 38.6%, p = 0.029. ELN risk groups were similarly distributed across cohorts (p = 0.194). A higher proportion of Black and Hispanic patients received intensive treatment compared to Whites: White 54% vs. Black 69% (p = 0.007), White vs. Hispanic 75.2% (p < 0.001). When data was restricted to 2020–2024 to align with the period of data collection for White patients (including only those with known treatment intensity), a smaller proportion of Hispanic (30/43, 69.8%) and Black patients (32/50, 64%) received intensive therapy (all pair-wise p > 0.05). “7 + 3”-based regimens (± gemtuzumab ozogamicin or FLT3 inhibitors) predominated among patients receiving frontline intensive therapy, accounting for 71/90 (78.9%) of Hispanics, 66/82 (80.5%) of Blacks, and 123/228 (54%) of Whites. Within the 2020–2024 sub-cohort, utilization of 7 + 3-based regimen remained high, at 19/30 (63.4%) for Hispanic and 24/32 (75%) for Black patients. Across all three groups, most remaining intensive therapies were CPX-351 or cladribine-based inductions. Within lower-intensity approaches, hypomethylating agent (HMA) with venetoclax was most frequently used in 158/236 (67%) patients- Hispanic: 11/25 (44%), Black: 20/32 (62.5%), White: 123/168 (73.2%). Within the same group, 13/25 (52%) Hispanic, 10/32 (31.3%) Black, and 15/168 (8.9%) White patients received single agent HMA. Within the 2020–2024 sub-cohort, utilization of HMA and venetoclax was recorded in 10/13 (77%) for Hispanic and 13/18 (72.2%) for Black patients. Allogeneic HCT rates were similar among races: White 36.7%, Black 29.8%, Hispanic 38.8%, p = 0.280 (Table 1).
Mutations by race
In the overall cohort, the most common mutations (> 10% of patients) included DNMT3A (24.5%), NPM1 (22.8%), FLT3-ITD (17.5%), TET2 (17.1%), TP53 (15.6%), RUNX1 (14.7%), NRAS (14.3%), ASXL1 (13.4%), SRSF2 (13.4%), and IDH2 (12.6%). Mutation distribution was largely comparable between races after multiple-comparison adjustment. KRAS mutations were more common in Black (10.9%) compared to White (5.4%) and Hispanic (2%) patients (p = 0.022). NOTCH1 mutations were more prevalent in Black (8.9%) compared to White (0.5%) and Hispanic (1%) patients (p < 0.001). However, all patients with NOTCH1 mutations had a variant allele frequency (VAF) of around 45–55%, and hence the possibility of these variants being a benign single-nucleotide polymorphism exists. ETV6 mutations were observed more frequently in Black (2.9%) and Hispanic (3.9%) patients than in White patients (0.74%) (p = 0.026). Similarly, CSF3R mutations were more common among Black (3%) compared with White (0.3%) and Hispanic (1%) patients (p = 0.029). However, the absolute number of cases harboring these mutations was small (ETV6, n = 10; CSF3R, n = 5), limiting the ability to draw definitive conclusions (Table 2; Fig. 1).
Table 2.
Mutations by race
| Mutations | Full cohort (N = 645) |
White (N = 413) |
Black (N = 114) |
Hispanic (N = 118) |
p (White vs. Black) | p (White vs. Hispanic) | p (Black vs. Hispanic) |
|---|---|---|---|---|---|---|---|
| DNMT3A | 151 (24.5%) | 99 (24.3%) | 27 (25.5%) | 25 (24.5%) | 1.000 | 1.000 | 1.000 |
| NPM1 | 145 (22.8%) | 98 (24.0%) | 17 (15.2%) | 30 (25.9%) | 0.101 | 0.765 | 0.101 |
| FLT3-ITD | 113 (17.5%) | 69 (16.7%) | 17 (14.9%) | 27 (22.9%) | 0.752 | 0.251 | 0.251 |
| TET2 | 105 (17.1%) | 68 (16.7%) | 19 (18.1%) | 18 (17.5%) | 1.000 | 1.000 | 1.000 |
| TP53 | 96 (15.6%) | 69 (16.9%) | 17 (15.9%) | 10 (9.8%) | 0.915 | 0.315 | 0.404 |
| RUNX1 | 90 (14.7%) | 59 (14.5%) | 17 (16.2%) | 14 (13.7%) | 0.968 | 0.968 | 0.968 |
| NRAS | 88 (14.3%) | 57 (14.0%) | 11 (10.5%) | 20 (19.6%) | 0.430 | 0.313 | 0.299 |
| ASXL1 | 82 (13.4%) | 52 (12.8%) | 15 (14.3%) | 15 (14.7%) | 1.000 | 1.000 | 1.000 |
| SRSF2 | 82 (13.4%) | 62 (15.3%) | 12 (11.4%) | 8 (7.84%) | 0.524 | 0.219 | 0.524 |
| IDH2 | 78 (12.6%) | 49 (12.0%) | 15 (14.2%) | 14 (13.5%) | 1.000 | 1.000 | 1.000 |
| FLT3-TKD | 51 (8.04%) | 26 (6.30%) | 11 (9.91%) | 14 (12.7%) | 0.400 | 0.120 | 0.654 |
| BCOR | 44 (7.22%) | 25 (6.14%) | 12 (11.9%) | 7 (6.93%) | 0.229 | 0.950 | 0.502 |
| PTPN11 | 44 (7.21%) | 26 (6.39%) | 10 (9.80%) | 8 (7.92%) | 0.822 | 0.822 | 0.822 |
| IDH1 | 39 (6.29%) | 26 (6.34%) | 8 (7.55%) | 5 (4.81%) | 0.821 | 0.821 | 0.821 |
| KRAS | 35 (5.75%) | 22 (5.41%) | 11 (10.9%) | 2 (1.98%) | 0.113 | 0.193 | 0.065 |
| STAG2 | 35 (5.75%) | 26 (6.39%) | 2 (1.98%) | 7 (6.93%) | 0.255 | 1.000 | 0.255 |
| U2AF1 | 31 (5.04%) | 25 (6.14%) | 3 (2.83%) | 3 (2.94%) | 0.458 | 0.458 | 1.000 |
| WT1 | 30 (4.91%) | 19 (4.67%) | 5 (4.90%) | 6 (5.88%) | 1.000 | 1.000 | 1.000 |
| SF3B1 | 27 (4.40%) | 18 (4.42%) | 5 (4.76%) | 4 (3.92%) | 1.000 | 1.000 | 1.000 |
| JAK2 | 22 (3.58%) | 14 (3.44%) | 4 (3.81%) | 4 (3.92%) | 1.000 | 1.000 | 1.000 |
| PHF6 | 21 (3.43%) | 16 (3.93%) | 3 (2.86%) | 2 (1.98%) | 1.000 | 1.000 | 1.000 |
| CBL | 18 (2.93%) | 9 (2.21%) | 5 (4.76%) | 4 (3.92%) | 0.457 | 0.457 | 1.000 |
| Biallelic CEBPA | 16 (2.56%) | 11 (2.70%) | 1 (0.92%) | 4 (3.70%) | 0.529 | 0.529 | 0.529 |
| BCORL1 | 15 (2.46%) | 9 (2.21%) | 5 (4.95%) | 1 (0.99%) | 0.318 | 0.695 | 0.318 |
| GATA2 | 14 (2.30%) | 7 (1.72%) | 2 (1.98%) | 5 (4.95%) | 1.000 | 0.206 | 0.667 |
| ZRSR2 | 14 (2.28%) | 7 (1.72%) | 5 (4.76%) | 2 (1.96%) | 0.231 | 1.000 | 0.668 |
| EZH2 | 13 (2.12%) | 8 (1.97%) | 2 (1.90%) | 3 (2.94%) | 1.000 | 1.000 | 1.000 |
| NOTCH1 | 12 (1.97%) | 2 (0.49%) | 9 (8.91%) | 1 (0.99%) | < 0.001 | 0.486 | 0.035 |
| SETBP1 | 12 (1.95%) | 7 (1.72%) | 3 (2.86%) | 2 (1.96%) | 1.000 | 1.000 | 1.000 |
| ETV6 | 10 (1.63%) | 3 (0.74%) | 3 (2.86%) | 4 (3.92%) | 0.156 | 0.098 | 0.719 |
| CUX1 | 9 (1.48%) | 6 (1.47%) | 1 (0.99%) | 2 (2.02%) | 1.000 | 0.987 | 0.987 |
| CSF3R | 5 (0.82%) | 1 (0.25%) | 3 (2.97%) | 1 (0.99%) | 0.079 | 0.538 | 0.621 |
Fig. 1.
Mutational landscape of AML patients included in the study. Oncoprint heatmap of gene mutations by race
Survival outcomes
Full cohort
The median duration of follow-up for the entire cohort was 3.3 years (95% CI, 3.2–3.5). The median OS after AML diagnosis for Whites was 2.2 years (95% CI, 1.7–4.0), Blacks was 1.5 years (95% CI, 1.3–2.3), and Hispanics was 2.4 years (95% CI, 1.8-not estimable) (log rank p = 0.045) (Fig. 2). No significant OS difference existed for Whites vs. Blacks (p = 0.097) or Hispanics (p = 0.26); however, Hispanics survived longer than Blacks (p = 0.011) (Supplemental Fig. 1a-1c).
Fig. 2.
Survival outcome of AML patients. Kaplan-Meier estimates of OS for entire cohort illustrated by race
There was a significant difference in median OS based on ECOG, prior myeloid malignancy, ELN risk, receipt of allogeneic HCT, and induction treatment intensity (all with log rank p < 0.00001). The details of median, 3-year, and 5-year OS for each group are presented in Table 3.
Table 3.
Overall survival (OS) by groups
| Groups | Level | Median OS, years (95% CI) | 3-year OS, years (95% CI) | 5-year OS, years (95% CI) | Log rank p-value |
|---|---|---|---|---|---|
| Full cohort | 2.1 (1.8–2.6) | 0.45 (0.41–0.49) | 0.36 (0.32–0.42) | ||
| Race | White | 2.2 (1.7–4.0) | 0.48 (0.43–0.53) | 0.39 (0.32–0.47) | 0.045 |
| Black | 1.5 (1.3–2.3) | 0.33 (0.25–0.44) | 0.27 (0.19–0.38) | ||
| Hispanic | 2.4 (1.8 - N.E.) | 0.48 (0.39–0.59) | 0.40 (0.31–0.51) | ||
| ECOG | 0–1 | 2.3 (1.9–3.4) | 0.47 (0.43–0.51) | 0.38 (0.33–0.44) | < 0.001 |
| ≥ 2 | 0.7 (0.3–1.2) | 0.24 (0.14–0.40) | 0.13 (0.05–0.37) | ||
| Sex | Female | 2.4 (1.8–4.8) | 0.47 (0.42–0.54) | 0.40 (0.33–0.48) | 0.065 |
| Male | 1.9 (1.5–2.4) | 0.43 (0.38–0.49) | 0.33 (0.26–0.40) | ||
| Prior malignancy | Yes | 1.0 (0.7–1.9) | 0.29 (0.20–0.42) | 0.26 (0.17–0.40) | < 0.001 |
| No | 2.3 (2.0–3.5) | 0.48 (0.44–0.52) | 0.38 (0.33–0.44) | ||
| Prior myeloid malignancy | Yes | 0.9 (0.7–1.2) | 0.26 (0.19–0.35) | 0.23 (0.17–0.32) | < 0.001 |
| No | 3.4 (2.2–4.3) | 0.51 (0.46–0.56) | 0.40 (0.35–0.47) | ||
| ELN risk | Favorable | N.E. (5.0 - N.E.) | 0.70 (0.62–0.79) | 0.59 (0.47–0.73) | < 0.001 |
| Intermediate | 3.3 (2.2–4.5) | 0.52 (0.44–0.61) | 0.35 (0.25–0.48) | ||
| Adverse | 1.1 (1.0–1.4) | 0.32 (0.27–0.38) | 0.28 (0.23–0.34) | ||
| Allogeneic HCT at CR1 | Yes | 8.2 (8.2 - N.E.) | 0.75 (0.69–0.81) | 0.63 (0.56–0.72) | < 0.001 |
| No | 1.0 (0.9–1.3) | 0.25 (0.21–0.31) | 0.17 (0.11–0.25) | ||
| Treatment intensity | Low intensity | 1.2 (1.0–1.6) | 0.30 (0.23–0.37) | 0.23 (0.16–0.32) | < 0.001 |
| Intensive | 4.3 (3.3 - N.E.) | 0.56 (0.51–0.61) | 0.46 (0.40–0.53) |
Univariate analysis
Besides race, OS was associated with age at diagnosis (Hazard ratio [HR] 1.04, p < 0.001), ECOG (≥ 2: HR 2.26, p < 0.001), prior myeloid malignancy (No: HR 0.47, p < 0.001), BM blast (HR 0.99, p = 0.016), reticulin fibrosis (grade 1: HR 1.39, grade 2: HR 1.80, grade 3: HR 5.55; overall p < 0.001), ANC (HR 1.01, p = 0.006), ELN risk (vs. favorable; intermediate: HR 1.94, adverse: HR 3.46; overall p < 0.001), allogeneic HCT (No: HR 4.56, p < 0.001), and induction treatment intensity (Lower intensity: HR 2.18, p < 0.001) (Supplemental Table 1).
Mutations in the following genes demonstrated association with OS (all adjusted p-value): TP53 (HR 4.02, p = 0.003), TET2 (HR 1.61, p = 0.013), ASXL1 (HR 1.79, p = 0.003), NPM1 (HR 0.47, p = 0.003), and U2AF1 (HR 2.21, p = 0.003) (Supplemental Table 2).
Multivariable analysis
The following variables were included in a full model: race, sex, age at diagnosis, ECOG, prior myeloid malignancy, ELN risk, receipt of allogeneic HCT, treatment intensity, and gene mutations (TP53, NPM1, FLT3-ITD, IDH1, IDH2, myelodysplasia-related). Race did not emerge as an independent predictor of OS. No prior myeloid malignancy (HR 0.72, p = 0.022), ELN risk (intermediate: HR 2.67, adverse: HR 2.66, both p < 0.001), no allogeneic HCT (HR 4.05, p < 0.001), TP53 mutation (HR 2.83, p < 0.001), and IDH2 mutation (HR 0.64, p = 0.028) were independent predictors of OS (Table 4).
Table 4.
Multivariate analysis (full model) affecting overall survival
| Variables | Level | HR | 95% CI | p-value |
|---|---|---|---|---|
| Race | White | Reference | ||
| Black | 1.29 | 0.92–1.80 | 0.14 | |
| Hispanic | 1.08 | 0.78–1.51 | 0.6 | |
| Sex | Female | Reference | ||
| Male | 1.25 | 0.98–1.60 | 0.07 | |
| Age at diagnosis | 1.01 | 1.00–1.02 | 0.069 | |
| ECOG | 0–1 | Reference | ||
| ≥ 2 | 1.40 | 0.93–2.09 | 0.11 | |
| Prior myeloid malignancy | Yes | Reference | ||
| No | 0.72 | 0.55–0.95 | 0.022 | |
| ELN risk | Favorable | Reference | ||
| Adverse | 2.66 | 1.63–4.33 | < 0.001 | |
| Intermediate | 2.67 | 1.64–4.35 | < 0.001 | |
| Allogeneic HCT at CR1 | Yes | Reference | ||
| No | 4.05 | 2.96–5.55 | < 0.001 | |
| Treatment intensity | Intensive | Reference | ||
| Lower intensity | 0.96 | 0.70–1.31 | 0.8 | |
| TP53 | Wt | Reference | ||
| Mut | 2.83 | 1.97–4.08 | < 0.001 | |
| NPM1 | Wt | Reference | ||
| Mut | 1.09 | 0.71–1.66 | 0.7 | |
| FLT3-ITD | Wt | Reference | ||
| Mut | 0.92 | 0.60–1.43 | 0.7 | |
| IDH1 | Wt | Reference | ||
| Mut | 0.73 | 0.41–1.31 | 0.3 | |
| IDH2 | Wt | Reference | ||
| Mut | 0.64 | 0.43–0.95 | 0.028 | |
| MR mutationsa | Wt | Reference | ||
| Mut | 1.05 | 0.76–1.44 | 0.8 |
aMyelodysplasia-related (MR) mutations include SF3B1, SRSF2, ASXL1, BCOR, U2AF1, STAG2, EZH2, SRSR2, ASXL1, RUNX1
A stepwise reduced model using AIC optimization yielded a comparable fit and retained race, sex, age at diagnosis, ECOG, prior myeloid malignancy, ELN risk, allogeneic HCT, and TP53 and IDH2 mutations. Consistent with the full model, race did not independently predict OS, while the same clinical and molecular factors remained significantly associated with outcomes, with effect estimates comparable to those observed in the full model. While not significant in the full model, age showed significant association in the reduced model (Supplemental Table 3).
White patient cohort
Univariate analysis
OS was associated with age at diagnosis (HR 1.04, p < 0.001), sex (Male: HR 1.33, p = 0.044), ECOG (≥ 2: HR 1.98, p = 0.009), prior myeloid malignancy (No: HR 0.50, p < 0.001), reticulin fibrosis (grade 1: HR 1.31, grade 2: HR 1.63, grade 3: HR 4.74; overall p < 0.001), ANC (HR 1.05, p < 0.001), ELN risk (vs. favorable; intermediate: HR 1.76, adverse: HR 3.25; overall p < 0.001), allogeneic HCT (No: HR 4.41, p < 0.001), and treatment intensity (lower intensity: HR 2.14, p < 0.001) (Supplemental Table 4).
Mutations in the following genes demonstrated association with OS (all adjusted p-value): TP53 (HR 4.48, p = 0.003), ASXL1 (HR 1.83, p = 0.042), NPM1 (HR 0.46, p = 0.003), and U2AF1 (HR 2.38, p = 0.005) (Supplemental Table 5).
Multivariable analysis
Factors associated with worse OS in the full model included ELN intermediate- and adverse-risks (HR 2.55, p = 0.004 and 2.59, p = 0.003 respectively), lack of allogeneic HCT (HR 4.37, p < 0.001) and TP53 mutation (HR 3.52, p < 0.001), while IDH2 mutation (HR 0.55, p = 0.028) was associated with improved OS. The stepwise reduced model, which excluded sex, NPM1, FLT3-ITD, and MDS-related mutations, yielded highly similar results. ELN risk remained strongly prognostic (intermediate: HR 2.36, p = 0.001; adverse: HR 2.61, p < 0.001), along with lack of allogeneic HCT (HR 4.33, p < 0.001), TP53 mutation (HR 3.23, p < 0.001), and IDH2 mutation (HR 0.56, p = 0.028) (Supplemental Tables 6 & 7).
Black patient cohort
Univariate analysis
OS was associated with age at diagnosis (HR 1.04, p < 0.001), prior myeloid malignancy (No: HR 0.4, p < 0.001), BM blast percent (HR 0.99, p = 0.009), reticulin fibrosis (grade 1: HR 1.71, grade 2: HR 2.59, grade 3: HR 25.21; overall p < 0.001), ELN risk (vs. favorable; intermediate: HR 5.10, adverse: HR 5.84; overall p = 0.003), and allogeneic HCT (No: HR 3.75, p < 0.001) (Supplemental Table 8).
Many gene mutations were excluded from univariate analysis due to small number of events. Among those included, only TP53 mutation (HR 3.83, p = 0.001) was associated with OS (Supplemental Table 9).
Multivariable analysis
The full model, using the same candidate variables as the main analysis, identified ECOG (≥ 2: HR 3.15, p = 0.01), ELN intermediate risk (HR 6.13, p = 0.049), and lack of allogeneic HCT (HR 3.48, p = 0.003) as independent predictors of inferior OS. In the reduced model, which retained ECOG, prior myeloid malignancy, ELN risk, allogeneic HCT, and TP53 mutation, ECOG (≥ 2: HR 2.82, p = 0.013), ELN risk (intermediate: HR 6.86, p = 0.011; adverse: HR 6.24, p = 0.014), TP53 mutation (HR 2.24, p = 0.042) and lack of HCT (HR 3.89, p < 0.001) remained significant (Supplemental Tables 10 & 11).
Hispanic patient cohort
Univariate analysis
OS was associated with age at diagnosis (HR 1.05, p < 0.001), ECOG (≥ 2: HR 3.85, p < 0.001), prior myeloid malignancy (No: HR 0.48, p = 0.008), aspirate cellularity (HR 0.99, p = 0.017), reticulin fibrosis (grade 3: HR 7.63, p < 0.001), ELN risk (vs. favorable; intermediate: HR 1.18, adverse: HR 2.90; overall p = 0.001), allogeneic HCT (No: HR 6.12, p < 0.001), and treatment intensity (low intensity: HR 5.51, p < 0.001) (Supplemental Table 12).
Many gene mutations were excluded from univariate analysis due to small number of events. None among those included in the model showed association with OS (Supplemental Table 13).
Multivariable analysis
The full model demonstrated that adverse-risk ELN (HR 4.42, p = 0.015), lack of allogeneic HCT (HR 5.81, p < 0.001), and IDH1 mutation (HR 5.11, p = 0.016) were independently associated with OS. The reduced model, which included ECOG, prior myeloid malignancy, ELN risk, allogeneic HCT, treatment intensity, and IDH1 mutation, showed similar but more granular associations. Adverse ELN risk (HR 2.67, p = 0.025), lack of allogeneic HCT (HR 5.92, p < 0.001), and IDH1 mutation (HR 5.52, p = 0.003) were associated with inferior OS (Supplemental Tables 14 & 15).
OS by race among ELN risk groups
OS was evaluated across racial groups within the 2022 ELN risk strata. Among patients with favorable-risk disease, no significant differences in OS were observed across races (log-rank p = 0.69). In intermediate-risk disease, Blacks had inferior OS to Whites (p = 0.02) and Hispanics (p = 0.003), while OS did not differ between Whites and Hispanics (p = 0.27). No significant racial differences in OS were observed within the adverse-risk group (log-rank p = 0.56) (Figs. 3, 4 and 5).
Fig. 3.
Survival outcome of ELN favorable risk group. Kaplan-Meier estimates of OS illustrated by race
Fig. 4.
Survival outcome of ELN intermediate risk group. Kaplan-Meier estimates of OS illustrated by race
Fig. 5.
Survival outcome of ELN adverse risk group. Kaplan-Meier estimates of OS illustrated by race
OS by race according to treatment intensity
OS was also analyzed by treatment intensity. Among patients receiving intensive induction, Blacks had significantly worse OS compared with Whites (p = 0.001) and Hispanics (p < 0.001), whereas survival was similar between Whites and Hispanics (p = 0.45) (Fig. 6). Within the lower-intensity group, Hispanics demonstrated inferior OS compared to Whites (p = 0.01) but similar to Blacks (p = 0.08). OS did not differ between Whites and Blacks (p = 0.77) (Fig. 7a). Restricting the analysis to HMA/venetoclax showed comparable OS between Whites and Blacks (p = 0.84). Compared to Hispanics, the other two cohorts had better OS, albeit significant only for Blacks (p vs. Black, 0.04; vs. White, 0.06) (Fig. 7b).
Fig. 6.
Survival outcome of AML patients based on treatment intensity. Kaplan-Meier estimates of OS for patients treated with intensive chemotherapy illustrated by race
Fig. 7.
Survival outcome of AML patients treated with lower intensity treatment. (a) Kaplan-Meier estimates of OS for patients treated with all lower intensity chemotherapy illustrated by race, (b) Kaplan-Meier estimates of OS for patients treated with hypomethylating agents/venetoclax illustrated by race
In each of the races, patients treated with intensive induction had better OS compared to lower-intensity, but the difference was not statistically significant in Black population (p < 0.0001 for both White and Hispanic, p = 0.19 for Black) (Supplemental Fig. 2a-2c).
Survival outcomes within the 2020–2024 cohort
A total of 520 patients diagnosed between 2020 and 2024 were evaluated, including 53 Black and 45 Hispanic patients. Baseline characteristics and mutational distributions were well-balanced across races, with no statistically significant differences observed. In the OS analysis, race did not emerge as an independent prognostic factor. Comprehensive baseline data, race-stratified multivariable models, and corresponding survival curves are detailed in Supplementary Tables 16–21 and Supplementary Figs. 3–4.
Discussion
This study offers a real-world evaluation of racial differences in AML, integrating genomic, treatment, and survival data from a large institutional cohort. By directly comparing White, Black, and Hispanic patients within a single tertiary center, our analysis minimizes variability in treatment access and supportive care and allows a more refined assessment of biologic and clinical factors that may underlie reported disparities in outcomes.
We aimed to define biologic differences in the genomic landscape of AML across White, Black, and Hispanic patients, an area that is characterized with inconsistent findings in the existing literature. In our institutional cohort, we observed a non-significant trend toward higher FLT3-TKD frequency among Black and Hispanics compared with White patients. Notably, KRAS mutations were significantly more frequent in Black patients, aligning with prior reports of RAS-pathway enrichment in younger Black AML populations [14]. We also observed a markedly higher prevalence of NOTCH1 variants in Black patients; however, the consistently high VAF raises the possibility that these represent germline polymorphisms rather than AML-specific driver events, underscoring the importance of cautious interpretation. Without germline validation or ancestry analysis, these NOTCH1 variants may also simply reflect geographically enriched benign variants or variants of uncertain significance within the local referral population.
In the Hispanic cohort, the most frequently observed mutations in our study (NPM1, DNMT3A, FLT3-ITD, and NRAS) were broadly consistent with prior reports, though the relative frequencies differed [19]. Compared with published non-Hispanic cohorts, both our data and prior studies suggest a lower incidence of several common mutations, including NPM1, DNMT3A, FLT3-ITD, and epigenetic regulators, which may reflect population-specific genetic background [19, 20]. Other studies have reported underrepresentation of IDH1 and enrichment of WT1 mutations among Hispanics, findings that were not uniformly replicated in our cohort, highlighting ongoing heterogeneity across datasets [16]. Despite subtle difference in mutational landscape, consistent with prior reports, we observed no significant differences in cytogenetic or ELN risk group distribution between Hispanic and White patients [15].
Among Black patients, prior studies, particularly in adolescent and young adults, have demonstrated enrichment of mutations like ASXL1, KRAS, ZRSR2, BCOR, IDH2 and CALR, while White patients were more likely to harbor NPM1 or biallelic CEBPA variants [13, 14]. In our cohort, Blacks showed numerically higher frequencies of ASXL1, ZRSR2, BCOR, and IDH2 mutations compared with Whites, though only KRAS reached statistical significance. These findings are largely concordant with existing literature and suggest that some previously reported molecular differences may be attenuated in older or more heterogeneous adult cohorts. Like prior studies, we found no significant racial differences in ELN cytogenetic or molecular risk stratification, indicating that while specific gene mutations may vary in frequency, overall genomic risk categorization remains comparable [13]. Differences across studies likely reflect varying cohort size, age distribution, sequencing platforms, ancestry admixture, and inclusion criteria, emphasizing the need for larger, uniformly profiled datasets to more definitively characterize race-associated genomic patterns.
In our institutional cohort, OS did not differ significantly between Black and White patients, a finding that contrasts with several large population-based and cooperative group studies. For example, Linet et al. observed a significantly increased hazard of death for Blacks compared to Whites with AML [21]. Similarly, Bhatnagar et al. demonstrated worse outcomes for Blacks across both younger and older age groups, with this disparity persisting after multivariable adjustment [13]. Inferior survival for Blacks has also been reported in clinical trial populations, including patients treated on CALGB/Alliance protocols, where Black race independently conferred a higher risk of death [13]. In contrast, a large registry study from England reported superior short- and long-term survival among Blacks compared with Whites, underscoring substantial geographic and health-system–specific variation [5]. We likewise observed no difference in OS between White and Hispanic patients. This finding is consistent with SEER-based analyses demonstrating similar adjusted survival between Whites and Hispanics across AML subgroups [21]. Other institutional studies have reported superior unadjusted survival among Hispanic patients, though this advantage was attenuated after adjustment for age, PS, and treatment site, suggesting confounding by demographic and clinical factors [16]. Conversely, some reports have noted worse age-adjusted survival among Hispanic patients [15]. Race was not an independent predictor of OS after multivariable adjustment in our study, likely reflecting the impact of uniform access to specialized care, contemporary treatment strategies, and consistent supportive management at a tertiary center, factors that may attenuate disparities driven by socioeconomic status or treatment availability in population-based studies. Overall, these findings suggest that when equivalent care is delivered, survival outcomes in AML may be comparable across races despite differences in molecular features.
No survival differences were observed among races within the ELN favorable or adverse risk categories. However, within ELN intermediate risk, both Whites and Hispanics demonstrated superior OS compared with Blacks. The intermediate-risk category represents a biologically heterogeneous group encompassing a wide range of cytogenetic abnormalities. In our cohort, normal karyotype was observed in 67% of White, 42% of Black, and 64% of Hispanic patients, with the remainder exhibiting trisomies, deletions, t(9;11), or other structural abnormalities. Prior studies have shown that certain abnormalities like isolated or predominant trisomy 8 are associated with inferior outcomes compared with normal-karyotype AML, with an absolute reduction in 5-year OS of 13.5%, even after adjustment for ELN 2022 risk [22]. These data underscore that grouping cytogenetically diverse entities under a single ELN intermediate-risk designation may obscure meaningful prognostic differences. Accordingly, the higher proportion of chromosomal abnormalities among Blacks in our intermediate-risk cohort may have contributed to the observed disparity in OS.
Within the lower-intensity cohort, Hispanic patients experienced worse OS compared with both Blacks and Whites, whose outcomes were comparable. In our study, the Hispanic population was relatively younger overall, a demographic characteristic typically associated with better PS and greater eligibility for intensive therapy. Consequently, Hispanics who received lower-intensity regimens likely represented a more selected subgroup with poorer baseline PS or greater comorbidity burden that precluded intensive treatment. This treatment selection bias may, in part, explain the inferior OS observed in the Hispanics treated with lower-intensity approaches. While Hispanic patients treated with the HMA and venetoclax combination showed significantly worse outcomes than Blacks and a non-significant trend toward inferior survival compared to Whites, these findings remain exploratory due to the small Hispanic sample size (n = 11).
After adjustment for relevant covariates, treatment intensity was not independently associated with OS within any race. This finding is consistent with a multicenter study comparing intensive versus lower-intensity therapy in AML [23]. In the retrospective cohort of that study, patients receiving lower-intensity treatment were older, had more comorbidities, more adverse cytogenetic features, and experienced higher mortality with inferior 2-year OS - patterns like those observed in our cohort. Importantly, in the prospective cohort, once age and PS were accounted for, mortality risks were comparable between treatment intensities. Collectively, these findings suggest that patient-related factors rather than treatment intensity per se may primarily drive survival differences; however, a randomized clinical trial would be required to definitively determine comparative efficacy.
Our study has certain limitations. We acknowledge that racial disparities in healthcare access could be a source of bias. For context, a retrospective cohort study conducted at MCC involving patients with five non-leukemia malignancies (2008–2020) showed that black patients experienced longer intervals between diagnosis and presentation to MCC from community, while Hispanic patients were more likely to present with advanced-stage disease and delayed presentation after first-course treatment [24]. While that study did not identify the root causes of these disparities, systemic barriers such as limited transportation, insurance gaps, and other socioeconomic or geographic barriers likely played a major role. Furthermore, we must note that our own cohort analysis could be influenced by other unmeasured variables like donor availability and treatment adherence. Retrospective design results in selection bias, especially with exclusion of incomplete cases from multivariable models. This study includes single-center cohort from a tertiary referral institution and represents another source of bias. While the overall cohort was large, the small sample sizes in certain subgroups, such as Black and Hispanics receiving HMA and venetoclax, and the low frequency of specific mutations limited the statistical power for granular comparisons. Race is widely recognized as a social construct rather than a precise genetic category; consequently, significant intra-group genetic variation may exist that is not captured by these broad classifications, which would be limitation of our study in the absence of ancestral information.
Conclusion
Race is not an independent predictor of overall survival in AML. Instead, outcomes are primarily driven by ELN risk, specific mutations such as TP53 and IDH1/2, and allogeneic HCT. The inferior outcomes observed in Black patients with intermediate-risk disease and Hispanic patients receiving lower-intensity therapy suggest that biological heterogeneity and treatment selection bias may influence subgroup-specific survival.
Supplementary Information
Below is the link to the electronic supplementary material.
Author contributions
U.J., M.H, and O.C. conceptualized the study. U.J., F.T.Z., P.P., N.P., N.A.A., J.S., M.H., and O.C. did the data collection. V.A. and Z.T. performed the data analysis. U.J., Z.T., and O.C. prepared the final figures and tables. U.J. and O.C. wrote the manuscript with input from R.S., Z.X., S.Y., A.K., E.P., A.W., D.S., R.K., J.L., and M.H. All authors approved the final version of the manuscript.
Funding
None.
Data availability
For original data, please contact Onyee.Chan@moffitt.org.
Declarations
Consent to participate
Not applicable.
Competing interests
David Sallman- Consultancy: AbbVie, Agios, Debiopharm, Janssen, Johnson & Johnson, Molecular Partners, Novartis. Advisory Board: AbbVie, Agios, Astellas, AvenCell, Bristol-Myers Squibb, BlueBird Bio, Dark Blue Therapeutics, Geron, Novartis, Shattuck Labs, Servier, Syndax, Syros, Taiho Oncology. Research Funding: Aprea, Jaz. Andrew Kuykendall- Consultancy/Honoraria: Abbvie, Blueprint Medicines, Celgene/BMS, CTI Biopharma, DISC Medicine, Geron, GSK, Incyte, Karyopharm, MorphoSys, Novartis, PharmaEssentia, Protagonist, Silence Therapeutics. Research Funding: AJAX, Blueprint Medicines, BMS, Geron, GSK, Janssen, Kartos, MorphoSys, Novartis, Protagonist. Rami Komrokji- Abbvie: consulting. BMS: Research Grant, Advisory board. DSI: Advisory board and speaker bureau. Geron: Consultancy. GSK: Advisory board. Takeda- steering committee, consultancy. Pharma Essentia: Speaker Bureau, Advisory board. Rigel: Speaker Bureau, Advisory board. Novartis: Advisory board. Servier: Speaker Bureau, Advisory board. Sobi: Speaker Bureau, Advisory board. Stem line: Advisory board. Sumitomo Pharma: Advisory board. Jeffrey Lancet- Stemline Therapeutics – Consultant. GLG Consulting – Consultant. Biomea Fusion, Inc. – Research Funding (to institution). Jasper Therapeutics, Inc. – Research Funding (to institution). Onyee Chan- has served in a consulting or advisory role for Bristol Myers Squibb, Novartis, Servier, Syndax, and Takeda; has received honoraria from AbbVie, Aptitude Health, and OncLive; has received research funding from AbbVie, Cullinan Florentine, Daiichi-Sankyo, Jazz Pharmaceuticals, Kumquat, and Remix Therapeutics. Rest of the authors do not have any competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Mohammad Hussaini and Onyee Chan are both senior authors and contributed equally to this work.
References
- 1.Döhner H, Weisdorf DJ, Bloomfield CD (2015) Acute myeloid leukemia. N Engl J Med 373(12):1136–52 [DOI] [PubMed] [Google Scholar]
- 2.National Cancer Institute Cancer Stat Facts: Acute Myeloid Leukemia (AML). SEER. Available from: https://seer.cancer.gov/statfacts/html/amyl.html. Accessed 2025 Nov 24
- 3.Dhakal P, Loeffler BT, Ravindra A, Acharya L, Pyakuryal A, Bhatt VR (2024) Improved overall survival in acute myeloid leukemia over the last 15 years. J Clin Oncol 42(16_suppl):6543–6543 [DOI] [PubMed] [Google Scholar]
- 4.Shallis RM, Wang R, Davidoff A, Ma X, Zeidan AM (2019) Epidemiology of acute myeloid leukemia: recent progress and enduring challenges. Blood Rev 36:70–87 [DOI] [PubMed] [Google Scholar]
- 5.Liu H, Stanworth SJ, McPhail S, Bishton M, Rous B, Bacon A et al (2024) Impact of patient demographics on treatment outcomes in AML: a population-based registry in England, 2013–2020. Blood Adv 8(17):4593–605 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Han HJ, Choi K, Suh HS (2024) Impact of aging on acute myeloid leukemia epidemiology and survival outcomes: a real-world, population-based longitudinal cohort study. PLoS One 19(5):e0300637 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tawfik B, Pardee TS, Isom S, Sliesoraitis S, Winter A, Lawrence J et al (2016) Comorbidity, age, and mortality among adults treated intensively for acute myeloid leukemia (AML). J Geriatr Oncol 7(1):24–31 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dhakal P, Lyden E, Joshi U, Pyakuryal A, Loh KP, Klepin H et al (2023) Comorbidity burden and outcomes of older adults with acute promyelocytic leukemia: a National Cancer Database analysis of 2221 patients. Leuk Lymphoma 64(3):691–7 [DOI] [PubMed] [Google Scholar]
- 9.Nielsen LH, Kristensen DT, Jakobsen LH, Bøgsted M, Gregersen H, Madsen J et al (2024) Socioeconomic status and overall survival among patients with hematological malignant neoplasms. JAMA Netw Open 7(3):e241112 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hoff FW, Blum WG, Huang Y, Welkie RL, Swords RT, Traer E et al (2024) Beat-AML 2024 ELN–refined risk stratification for older adults with newly diagnosed AML given lower-intensity therapy. Blood Adv 8(20):5297–305 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Döhner H, Wei AH, Appelbaum FR, Craddock C, DiNardo CD, Dombret H et al (2022) Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood 140(12):1345–77 [DOI] [PubMed] [Google Scholar]
- 12.Patel MI, Ma Y, Mitchell BS, Rhoads KF (2015) Age and genetics. Am J Clin Oncol 38(2):159–64 [DOI] [PubMed] [Google Scholar]
- 13.Bhatnagar B, Kohlschmidt J, Mrózek K, Zhao Q, Fisher JL, Nicolet D et al (2021) Poor survival and differential impact of genetic features of Black patients with acute myeloid leukemia. Cancer Discov 11(3):626–37 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Larkin KT, Nicolet D, Kelly BJ, Mrózek K, LaHaye S, Miller KE et al (2022) High early death rates, treatment resistance, and short survival of Black adolescents and young adults with AML. Blood Adv 6(19):5570–81 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Darbinyan K, Shastri A, Budhathoki A, Helbig D, Snyder R, Pradhan K et al (2017) Hispanic ethnicity is associated with younger age at presentation but worse survival in acute myeloid leukemia. Blood Adv 1(24):2120–3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bradley T, Kwon D, Monge J, Sekeres M, Chandhok N, Thomassen A et al (2022) Molecular characteristics and outcomes in Hispanic and non-Hispanic patients with acute myeloid leukemia. EJHaem 3(4):1231–40 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Premnath N, Khan A, Pandey M, Pandey U, Hoff FW, Chung S et al (2023) Shattering myths regarding Hispanic ethnicity and survival in acute myeloid leukemia: insights from the National Cancer Database. J Clin Oncol 41(16_suppl):7032–7032 [Google Scholar]
- 18.Moffitt Cancer Center (2026) About Our Catchment Area. Available from: https://www.moffitt.org/research-science/community-outreach-and-engagement/catchment-area/. Accessed May 28
- 19.Gonzalez-Lugo JD, Madanat YF, Zapata N, Nicolet D, Klein V, Foucar C et al (2024) Molecular features of acute myeloid leukemia in Hispanics: an analysis from the Marrow Consortium. Blood 144(Supplement 1):3801–3801 [Google Scholar]
- 20.Gomez-Arteaga A, Mencia-Trinchant N, Tovar AM, Carrillo IO, Laguna AG, Fuller ER et al (2020) Abstract C061: molecular landscape of acute myeloid leukemia in Mexico. Cancer Epidemiol Biomarkers Prev 29(6_Supplement_1):C061–C061 [Google Scholar]
- 21.Linet MS, Curtis RE, Schonfeld SJ, Vo JB, Morton LM, Dores GM (2024) Survival of adult AML patients treated with chemotherapy in the U.S. population by age, race and ethnicity, sex, calendar-year period, and AML subgroup, 2001–2019. EClinicalMedicine 71:102549 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kristensen DT, Brøndum RF, Knudsen M, Severinsen MT, Grubach L, Andersen MK et al (2025) Outcomes and prognostic impact of trisomy 8 in acute myeloid leukemia patients treated with intensive chemotherapy. Blood Cancer J 15(1):146 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sorror ML, Storer BE, Fathi AT, Brunner A, Gerds AT, Sekeres MA et al (2021) Multisite 11-year experience of less-intensive vs intensive therapies in acute myeloid leukemia. Blood 138(5):387–400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kirtane K, Zhao Y, Amorrortu RP, Fuzzell LN, Vadaparampil ST, Rollison DE (2023) Demographic disparities in receipt of care at a comprehensive cancer center. Cancer Med 12(12):13687–13700. 10.1002/cam45992 PubMed PMID: 37114585 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
For original data, please contact Onyee.Chan@moffitt.org.