Treatment in Clinical Trials Could Potentially Overcome the Disparity in Outcomes of Patients with Acute Myeloid Leukemia from Disadvantaged Neighborhoods

Introduction

Population-based studies from publicly available databases have shown a disparity in the outcomes of patients (pts) with acute myeloid leukemia (AML) based on their racial background. However, such studies are limited as several aspects of socioeconomic (SE) data, an independent prognostic factor, still need to be captured. Area Deprivation Index (ADI) is one of the most advanced SE tools, incorporating 17 SE factors to rank neighborhoods based on disadvantaged status (Powell et al. Forefront, 2023). This abstract, reports the largest cohort of pts with molecular, cytogenetic (CG) and ADI data treated in a single institution.

Methods

Adult pts with AML available self-reported race and treated at MD Anderson Cancer Center from 3/2013 to 3/2023 were included. ADI data was downloaded from https://www.neighborhoodatlas.medicine.wisc.edu/May 10,2024. The MatchIt package was used to perform nearest neighbor matching without replacement, using a generalized linear model to estimate propensity scores based on the following covariates (gender, age, ELN 2022 risk, treatment intensity, ADI State Rank, and ADI National Rank), and aiming to match each racial unit (Non-Hispanic Black, NHB or Hispanic, H) with up to 9 Non-Hispanic Whites (NHW). Fisher’s exact test and Wilcoxon rank sum test were used to compare two groups. The probabilities of OS were estimated using Kaplan-Meier method. Backward multivariable analysis was used to predict factors influencing OS.

Results

2442 pts were identified (NHW-2032, NHB-202, H-109, and Asians, A-99). Data on Asians will not be reported in this abstract. In the MVA on unmatched cohort (n=2343), age ≥60 (HR-1.4, 95%CI 1.2-1.6, p<0.001), complex CG (HR-1.88, 1.6-2.2, p<0.001), TP53 mutation (m) (HR-1.3, 1.1-1.5, p=0.001 led to shorter OS and favorable/intermediate ELN 2022 risk (HR-0.42/0.84, 0.4-0.5/0/7-0.97, p<0.001/=0.02), intensive treatment (HR-0.82, 0.7-0.95, p=0.01), denovo AML (vs. secondary, HR-0.58, 0.5-0.7, p<0.001), allogeneic stem cell transplant (SCT, HR-0.33, 0.3-0.4, p<0.001) led to longer OS. Racial differences (NHB vs. NHW, HR-1.12, 0.94-1.3, p=0.22/H vs. NHW, HR-0.79, 0.6-1.03, p=0.08) did not affect OS. In the matched cohort (NHB, n=201 vs. NHW, n=1206), the only difference was SCT rate (13% vs. 20%, p=0.04). Clinical trial enrollment was similar (73% vs. 76%, p=0.37). In MVA, favorable risk (HR-0.39, 0.3-0.5, p<0.001), denovo AML (HR-0.59, 0.51-0.69, p<0.001), clinical trial (HR-0.7, 0.6-0.8, p<0.001), SCT (HR-0.32, 0.3-0.4, p<0.001) led to longer OS. Whereas age ≥60 (HR-1.5, 1.3-1.7, p<0.001), complex CG (HR-2.13, 1.7-2.6, p<0.001), TP53m (HR-1.3, 1.03-1.6, p=0.02) adversely affected OS. Strikingly, race did not affect OS (HR-1.08, 0.9-1.3, p=0.43). The median OS was 11.8 vs. 10.8 months (mo), p=0.32. Similarly, in the H (n=109) vs. NHW (n=981) matched cohort, the following mutations were more common in NHW: ASXL1 (2% vs. 9%, p=0.04), DNMT3A (9% vs. 21%, p=0.01), WT1 (0 vs. 6%, p=0.01). Clinical trial enrollment (69% vs. 75%, p=0.21); SCT rate (19% vs. 25%, p=0.20) was similar. The MVA showed favorable/intermediate risk (HR-0.32/0.75, 0.2-0.4/0.6-0.9, p<0.001/=0.01), denovo AML (HR-0.52, 0.4-0.6, p<0.001), clinical trial (HR-0.72, 0.6-0.9, p<0.001), SCT (HR-0.34, 0.3-0.4, p<0.001) led to longer OS, and age ≥60 (HR-1.6, 1.3-1.9, p<0.001), complex CG (HR-1.78, 1.4-2.3, p<0.001), TP53m (HR-1.4, 1.06-1.9, p=0.02) led to shorter OS. Race did not affect OS (HR-0.8, 0.6-1.05, p=0.10). The median OS was 14.3 vs. 14.9 mo, p=0.43. In NHB vs. H, (87 each) sAML was more common in H (23% vs. 38%, p=0.05) and IDH1 was more frequent in NHB (10% vs. 1%, p=0.04). In the MVA, favorable risk (HR-0.44, 0.2-0.9, p=0.02), denovo AML (HR-0.43, 0.3-0.7, p=0.002, SCT (HR-0.3, 0.2-0.6, p<0.001) led to longer OS and age ≥60 (HR-2.5, 1.6-3.9, p<0.001), complex CG (HR-2.3, 1.4-4, p=0.002) led to shorter OS. Race did not affect OS (HR-0.8, 0.6-1.2, p=0.35). The median OS was 17 vs. 14.3 mo, p=0.85.

Conclusion

There is a disparity in the outcomes of pts with AML from disadvantaged neighborhoods (higher ADI rank) regardless of racial background. There is an unmet need to improve the outcomes in this subgroup of pts. Our study showed that clinical trial treatment and SCT could help overcome this disparity. We plan to identify the barriers to clinical trial participation as the next step to improve outcomes of pts from higher ADI ranks.