Molecular Features of Acute Myeloid Leukemia in Hispanics: An Analysis from the Marrow Consortium

Background: Outcomes of patients with acute myeloid leukemia (AML) depend on several patient-specific factors, such as age and performance status, as well as leukemia-specific characteristics, such as cytogenetic and molecular abnormalities at diagnosis. Additionally, emerging research suggests that race, ethnicity, and socioeconomic status may also impact patient outcomes. Hispanics are one of the largest minority populations, making up close to 20% of the US population. However, they continue to be underrepresented in clinical trials. Data on AML-associated driver mutations in Hispanic populations is scarce. According to small studies in different Hispanic populations, Hispanic patients may have different and more adverse AML-associated driver mutations, younger ages of presentation, and worse outcomes. Methods: We retrospectively analyzed clinical and molecular characteristics at diagnosis and outcomes of adult Hispanic patients in two cohorts. 1) United States (US) Cohort as part of the Myeloid Malignancy Association on Rapid Research Outcomes Working Group (MARROW) Consortium. 2) Mexico Cohort, patients from Hospital Universitario Jose Eleuterio Gonzalez in Monterrey and from Instituto Nacional de Cancerologia in Mexico City. We additionally used the non-Hispanic patients in the US Beat AML trial dataset (NCT03013998) for comparison of molecular characteristics. Results: Two hundred forty-one Hispanic patients were included, 159 from the US and 82 from Mexico (MEX). Median age at diagnosis was 52 years (y; range, 16-93), and it was significantly lower in the MEX cohort (53y US vs 44y MEX; p<0.001), and younger than median age at AML diagnosis. Half of the patients were females. The most prevalent gene mutations occurred at similar rates between US and MEX cohorts and were: FLT3-ITD in 15% (US 13% vs MEX 18%, p=0.25), NRAS in 12% (US 12% vs MEX 11%, p=1.00), and NPM1 in 11% of patients (US 12% vs MEX 10%; p=0.67). However, the cohort from MEX had more common mutations in IDH1 (overall, 5%; MEX 10% vs US 3%, p= 0.04), WT1 (6%; MEX 11% vs US 4%, p= 0.05), and DNMT3A (6%; MEX 11% vs US 4%, p= 0.05). Mutations in RUNX1 (7%) were the only ones occurring at significantly higher rates in the US cohort (US 9% vs MEX 1%, p= 0.01). There was also a trend toward a higher frequency of TP53 mutations (8% overall) in the US cohort ((US 11% vs MEX 4%, p= 0.08). When compared with the Beat AML non-Hispanic cohort, mutations that are less prevalent in our Hispanic cohort (MEX+US) were NPM1 (11% Hispanic vs 27% Beat ; p<0.001), DNMT3A (6% Hispanic vs 22% Beat ; p<0.001), FLT3-ITD (15% Hispanic vs 24% Beat; p=0.002), IDH2 (8% Hispanic vs 13% Beat; p=0.04), RUNX1 (7% Hispanic vs 12% Beat; p=0.02), SRSF2 (4% Hispanic vs 13% Beat; p< 0.001), U2AF1 (1% Hispanic vs 6% Beat ; p< 0.001), and TET2 (8% Hispanic vs 14% Beat; p= 0.03). Conclusion: This real-world analysis includes the largest reported Mexican dataset of AML patients with molecular data and shows that Hispanics have distinct AML-associated driver mutations and younger ages of presentation compared with non-Hispanic patients. NPM1 mutations are less prevalent in Hispanics in both the US and MEX cohorts, as are splicing mutations, TET2 and DNMT3A mutations, which could be related to the younger age of presentation. Furthermore, Mexican patients with AML have more common mutations in WT1 and IDH1 compared with Hispanic US patients. Genetic susceptibility or environmental factors could potentially explain these variations. This study underlines the importance of including Hispanic and other diverse populations in clinical trials and remarks on the genetic heterogeneity of the Hispanic populations, which is a cultural and not a genetically defined ethnicity. Our study is limited by its retrospective nature and limited sample size, and further research is needed to corroborate these data. We plan to have cytogenetic and outcomes data presented at the meeting.