Torres Aracelis Z, Phelan-Emrick Darcy, Castillo-Salgado Carlos
Department of Epidemiology, Bloomberg School of Public Health, Johns Hopkins, Baltimore, MD, United States.
Baltimore City Health Department, Baltimore, MD, United States.
Front Oncol. 2018 Oct 30;8:471. doi: 10.3389/fonc.2018.00471. eCollection 2018.
Though cancer research has traditionally centered on individual-level exposures, there is growing interest in the geography of both cancer and its risk factors. This geographic and epidemiological research has consistently shown that cancer outcomes and their known causal exposures exhibit geographic variation that coincide with area-level socioeconomic status and the composition of neighborhoods. A retrospective study was conducted to evaluate geospatial variation for female breast, cervical, and colorectal cancer incidence in Baltimore City. Using a Maryland Cancer Registry dataset of incident breast, cervical, and colorectal cancers ( = 4,966) among Baltimore City female residents diagnosed from 2000 to 2010, spatial and epidemiological analyses were conducted through choropleth maps, spatial cluster identification, and local Moran's I. Ordinary least squares regression models identified characteristics associated with the geospatial clusters. Each cancer type exhibited geographic variation across Baltimore City with the neighborhoods showing high incidence differing by cancer type. Specifically, breast cancer had significant low incidence in downtown Baltimore while cervical cancer had high incidence. The neighborhood covariates associated with the geographic variation also differed by cancer type while local Moran's I identified discordant clusters. Cancer incidence varied geographically by cancer type within a single city (county). Small area estimates are needed to detect local patterns of disease when developing health and preventative programs. Given the observed variability of community-level characteristics associated with each cancer type incidence, local information is essential for developing place-, social-, and outcome-specific interventions.
尽管癌症研究传统上集中在个体层面的暴露因素，但人们对癌症及其风险因素的地理分布越来越感兴趣。这种地理和流行病学研究一直表明，癌症结局及其已知的因果暴露呈现出与地区层面社会经济地位和社区构成相一致的地理差异。开展了一项回顾性研究，以评估巴尔的摩市女性乳腺癌、宫颈癌和结直肠癌发病率的地理空间差异。利用马里兰州癌症登记处2000年至2010年诊断的巴尔的摩市女性居民中乳腺癌、宫颈癌和结直肠癌（=4966例）的发病数据集，通过分级统计图、空间聚类识别和局部莫兰指数进行了空间和流行病学分析。普通最小二乘回归模型确定了与地理空间聚类相关的特征。每种癌症类型在巴尔的摩市都呈现出地理差异，不同癌症类型的高发病率社区也有所不同。具体而言，巴尔的摩市中心的乳腺癌发病率显著较低，而宫颈癌发病率较高。与地理差异相关的社区协变量也因癌症类型而异，同时局部莫兰指数识别出不一致的聚类。在一个单一城市（县）内，癌症发病率因癌症类型而异。在制定健康和预防计划时，需要进行小区域估计以检测疾病的局部模式。鉴于观察到的与每种癌症类型发病率相关的社区层面特征的变异性，本地信息对于制定针对地点、社会和结局的干预措施至关重要。
