AUTHORS
Reena Pabari, MD, MSc, FRCPC, Jim Wang2, Todd A. Alonzo, PhD3, Lena E. Winestone, MD, MPH4, Erin Marcotte, MPH, PhD5, Johann Hitzler, MD6, Robert B. Gerbing, Ph.D7, Todd Cooper I, DO8, Anupam Verma, MD9
AFFILIATIONS
1 Hospital for Sick Children, Toronto, Ontario, Canada
2 Children's Oncology Group, Monrovia, United States
3 Keck Medicine of USC, Monrovia, California, United States
4 UCSF Benioff Children's Hospital San Francisco, San Francisco, California, United States
5 University of Minnesota, Minneapolis, Minnesota, United States
6 Hospital for Sick Children, Toronto, Ontario, Canada
7 Children's Hospital of Philadelphia-Main Building, Philadelphia, Pennsylvania, United States
8 Seattle Children's Hospital, Seattle, Washington, United States
9 Rainbow Babies & Children Clinic, Cleveland, Ohio, United States
ABSTRACT
Introduction: Children with Down syndrome (DS) have an increased risk of developing myeloid leukemia (ML-DS). Patients with ML-DS were not included in prior analyses of non-DS acute myeloid leukemia (AML) that identified Black and Hispanic patients and those with lower socioeconomic status (SES) to have worse survival and increased treatment-related toxicity (Winestone 2019, Winestone 2023). Identifying inequities related to race, ethnicity, and SES in patients with ML-DS may inform targeted interventions and strategies to improve outcomes in this unique and vulnerable population.
Methods: We retrospectively evaluated patients with ML-DS enrolled on three Children’s Oncology Group Phase III trials (A2971, AAML0431, and AAML1531) for the impact of race, ethnicity, and proxies of SES (insurance type and geocoded indices) on event-free survival (EFS) and overall survival (OS). Only patients residing within the United States were included. Zipcode geocoding was performed in order to identify the patient-specific SES Social Vulnerability Index (SVI) quartile, a composite measure of housing, education, employment and poverty level, where quartile 1 represents those with the highest SES and lowest vulnerability, and quartile 4 encompasses those with lowest SES or highest vulnerability. Insurance type was categorized as private (including Military) or public (Medicare, Medicaid). Analyses were stratified by study and adjusted for relevant covariates including sex and disease risk group if known. Statistical methods include chi-square, Fisher’s exact, and nonparametric tests for group comparisons, with significance as p<0.05.
Results: We evaluated 117 patients from A2971 Arm B, 184 patients from AAML0431 and 144 patients from AAML1531 (108 standard risk and 36 high risk). There were more non-Hispanic patients enrolled on A2971 compared with AAML0431 and AAML1531 (87.9% vs 74.6% and 76.1%, p=0.016). In contrast to previously reported data from patients with non-DS AML, there were no significant differences in EFS based on race (non-Black vs Black; A2971: HR 1.55 [95% CI 0.6-4.1], AAML0431: HR 0.38 [0.1-2.9], AAML1531: HR 1.50 [0.5-4.6]), ethnicity (non-Hispanic White vs Hispanic; A2971: HR 1.05 [95% CI 0.3-3.6], AAML0431: HR 0.40 [0.1-1.8], AAML1531: HR 1.62 [0.5-5.1]), insurance type (public vs private; A2971: HR 0.97 [95% CI 0.2-5], AAML0431: HR 1.24 [0.5-3.3], AAML1531: HR 0.62 [0.2-1.7]) or SES SVI (quartile 1 vs 4; A2971: HR 1.18 [95% CI 0.4-3.9], AAML0431: HR 0.61 [0.2-2.4], AAML1531: HR 1.55 [0.5-5.1]). There were also no significant differences in OS based on race (non-Black vs Black; A2971: HR 1.53 [95% CI 0.5-4.6], AAML0431: HR 0.54 [0.1-4.2], AAML1531: HR 1.85 [0.6-5.9]), ethnicity (non-Hispanic White vs Hispanic; A2971: HR 1.46 [95% CI 0.4-5.2], AAML0431: HR 0.30 [0-2.4], AAML1531: HR 1.49 [0.4-5.3]), insurance type (public vs private; A2971: HR 1.19 [95% CI 0.2-6.5], AAML0431: HR 1.19 [0.3-4.1], AAML1531: HR 0.43 [0.1-1.3]), or SES SVI quartile (1 vs 4; A2971: HR 2.5 [95% CI 0.6-10.5], AAML0431: HR 0.58 [0.1-3], AAML1531: HR 1.31 [0.4-4.5]).
Conclusions: This is the largest report evaluating the impact of socioeconomic factors on outcomes in children with ML-DS. However, SES data collection across prior COG ML-DS trials was heterogeneous, limiting our ability to perform a combined analysis. On each of the three trials included in this report, outcomes in patients with ML-DS did not vary significantly by race, ethnicity, or SES. These results may be related to the unique biology of ML-DS, and variables which impact outcome in ML-DS may be distinct from those of non-DS AML. Given the known vulnerability of patients with ML-DS to infectious toxicity, next steps include evaluation of the association of race, ethnicity, and SES on the cumulative incidence of grade ≥3 serious infections. Our results highlight the importance of collection of socioeconomic data on clinical trials in order to identify factors impacting outcomes for children with ML-DS.
DISCLOSURES
Reena Pabari, MD, MSc, FRCPC: No relevant conflicts of interest to declare.