The myelodysplastic syndromes (MDS) are a heterogeneous group of hematopoietic disorders. This study aims to advance the use of spatial modeling in disease etiology and monitoring based on reports on a large population (n=984) of MDS patients diagnosed in the Eastern United States. The spatial MDS clustering was analyzed using SaTScan, and patient clinical characteristics were analyzed using logistic regression and Cox hazards models adjusting for covariates. One main and five secondary spatial clusters (p-value <10(-17)-10(-7)) were identified. Patients living in high vs. low MDS incidence clusters tended to be older (ORadj= 1.04 [1.004, 1.07]) and smokers (ORadj=2.9 [1.1, 7.4]). Mortality were associated with hemoglobin (HRadj=0.7 [0.5, 0.9]), neutrophils (HRadj=0.7 [0.6, 0.96]), platelets (HRadj=0.5 [0.4, 0.7]), and blast (HRadj=1.4 [1.1, 1.8]), but not clusters. The results suggest large geographic variations in MDS incidence rates. The biological aggressiveness of the disease is unlikely to be associated with its spatial distribution.