Comparison of demographics, disease characteristics, and outcomes between African American patients and White patients with myelodysplastic syndrome: A population-based study.

Abstract

7051 Background: Myelodysplastic syndrome (MDS) consists of a heterogeneous group of clonal myeloid neoplasms, classified as a malignancy since 2001. Previous studies have described inferior outcomes for African Americans (AA) versus Whites (W) in several solid tumors and hematological malignancies, including AML. We performed an analysis of the Survival, Epidemiology, and End Results (SEER) cancer registry to analyze baseline characteristics and various outcomes. Methods: Using SEER (18 Registries, 2000-2018, Nov 2020 submission), we included 37,564 patients (pts) with microscopically confirmed MDS, age > = 20, diagnosed between 2000-2013 based on ICO-O3 codes. MDS sub-types were grouped into low, intermediate, and high-risk disease. We compared baseline characteristics, mortality rates per attributed cause of death (COD) and overall survival (OS). Multivariate Cox regression and Propensity Score analyses were conducted to reduce effect of confounding due to imbalance in baseline covariates. Results: We included 34,543 W (92%) and 3021 AA (8%) pts; There were more males (58%) among W vs AA (49%). At diagnosis the median age was 71 and 76 in AA and W subjects, respectively. Low and high risk MDS comprised 22.8% and 14.5% of cases among AA and 21.5% and 16.3% among W pts, respectively. 66.8% of AA were living in large metropolitan areas vs 60.2% for W pts. mOS was 33 months for AA and 26 months for W; HR for OS comparing W vs AA using a univariate Cox PH model was: 0. 79 (0.76- 0.82), p < 0.001. In a multivariate Cox-PH model, HR for OS after adjusting for sex, age at diagnosis, histology, urban-rural continuum, income group was: 0.90 (95%CI 0.86-0.94), p < 0.001. HRs after stratified cox PH model and propensity score matching model for age, sex, histology and income group were 0.92 (0.88-0.96), p < 0.001 and 0.93(0.88-0.98), p = 0.01 respectively. Analysis by histology shows statistically significant association between race and OS for refractory anemia and MDS-unclassifiable while not significant for refractory anemia with ringed sideroblasts, refractory anemia with excess blasts and MDS with del(5q). Incidence rate ratio for AA vs W for death attributed to the two most frequent CODs were: MDS/Leukemia: 0.74 (0.69-0.81, p < 0.001); CVD: 1.07 (0.94-1.21, ns). Conclusions: In this large study of MDS cases reported to the SEER registry, we observed differences in age, sex, disease subtype, socio-economic factors, mortality rate per attributed COD, survival outcomes between AA and W. The finding of AA having better OS outcomes than W pts is unexpected and needs to be further investigated, specifically the contribution of the less well-defined low-grade risk forms of MDS. Although caution in the interpretation is necessary due to multiple limitations inherent to the SEER dataset, this data does raise intriguing questions for future study.