County-level analysis reveals two high-incidence regions and potential environmental contributors to myeloid malignancies in Colorado

Abstract Background: Myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) are hematologic malignancies that result from an accumulation of mutations in the myeloid pathway of hematopoiesis. These diseases are influenced by both intrinsic and extrinsic factors; environmental exposures with leukemogenic potential implicated in the development of myeloid-type mutations include tobacco, benzene, pesticides, and radiation. Thus, we sought to investigate the geographic distribution of MDS/AML in Colorado, as counties with disproportionate clusters could offer insight into known and unknown environmental influences on the development of myeloid disease. Methods: We used data from the Colorado Cancer Registry to identify the incidence of MDS and AML in each of the 64 counties in Colorado between 2012 and 2022. To account for known epidemiologic distributions of myeloid disease, sex, age, and race-adjusted incidence rates were calculated using direct standardization for each year based on U.S. Census county-level population estimates. County-level environmental and sociodemographic data were obtained from publicly available sources: tobacco and radon test results from the Colorado Department of Public Health, benzene exposure reports from the U.S. Environmental Protection Agency database, and social vulnerability index (SVI) scores from the U.S. Centers for Disease Control and Prevention. Logistic regression was used to assess for associations between these variables and our adjusted MDS/AML incidences. Results: For the entire state of Colorado, the median sex-adjusted rate of MDS/AML incidence per 100,000 is 18.38 (range 2.61-49.34), race-adjusted rate is 20.22 (3.58-444.77), and the age-adjusted rate is 171.99 (22.32-338.46). Cheyenne and Hinsdale counties demonstrated significantly elevated adjusted incidence of MDS/AML compared to the rest of Colorado: Cheyenne sex-adjusted rate per 100,000 is 49.34 (p = 0.002), age-adjusted rate is 220.88 (p = 0.009), and race-adjusted rate is 338.46 (p = 0.001); Hinsdale sex-adjusted rate is 38.62 (p = 0.034), race-adjusted rate is 444.77 (p = 5E-12), and age-adjusted rate is 158.63 (p = 9E-12). In Cheyenne county, 13% of adults report tobacco use, compared to 11% statewide. Both Cheyenne and Hinsdale counties had a high proportion of radon tests exceed 4 pCi/L (EPA action threshold) at 67% and 68% respectively, compared to 47% statewide. Benzene exposures were not increasingly reported in these counties compared to the average county, and no correlation was found between benzene levels and MDS/AML incidence overall. Similarly, we found no significant association between county-level SVI scores and adjusted incidence rates. Conclusions: After adjusting for sex, age, and race, geographic analysis in Colorado identified two counties with significantly elevated rates of MDS and AML – Cheyenne and Hinsdale. While these counties were also found to have increased radon exposure compared to average, and Cheyenne was found to have slightly increased tobacco use, no association was found with benzene exposure or SVI scores compared to the rest of Colorado. These findings suggest that environmental exposures contribute to regional clustering of myeloid malignancies, though interpretation is limited by the lack of granularity with county-specific data, errors that result from the law of small numbers, and potential confounders such as population mobility. Future directions include incorporating patient-level data, such as addresses and mutational profiles, to more precisely assess the relationship between environmental exposures and the development of myeloid disease.