甲状腺癌发病与人群暴露风险的时空关联分析---以中国杭州市为例

The main goal of this proposal is to explore a general methodology for studying the association between environmental exposures and thyroid cancer (TC) effect by means of spatiotemporal random field theory. This theory is the tool of choice for rigorously accounting for important spatiotemporal variations and multi-sourced uncertainties related to environmental exposures and population health effect. Within the framework of the random field theory, the Bayesian Maximum Entropy model neatly synthesizes various sources of physical and epidemiological knowledge (core and case-specific) into spatiotemporal analysis. Therefore, unlike technical statistics, this approach relies on the blending of substantive knowledge (physical, biological, medical, social etc.) with powerful mathematics and a coherent rationale. Given the well-founded fact that certain TC may be caused by environmental exposures (SES-over-diagnose, persistent organic pollutants) the significance of these exposures are assessed in terms of a criterion that is based on the joint stochastic representation of exposure and health effect (TC) distributions in a composite space-time domain. In view of this criterion, the strength and consistency of the “exposure-TC” association are evaluated on the basis of the TC predictions that the combined physic-epidemiologic analysis generates in space-time. The main features of the approach will be demonstrated and insight will be gained by a real case study involving TC incidence and various environmental exposures in Hangzhou，China. This study will demonstrate the advantages of the stochastic human exposure analysis in assessing the exposure-effect association, also emphasize the links between spatiotemporal models of physical systems and population health-effect distribution thus suggesting directions for improving the current understanding of quantitative “exposure-health effect” association, The outcomes will enrich spatial epidemiology and provide theory and references for spatial epidemiological research of Thyroid cancer, as well as for spatial epidemiological research of other non-communicable diseases.

本申请以时空随机场理论为基础，探究环境暴露风险与甲状腺癌症发生时空关联的通用方法。该方法在研究环境暴露及其响应之间关系时，能兼顾时空分布变化及其不确定性。在随机场理论框架下，贝叶斯最大熵模型巧妙地综合来自多种渠道的物理和流行病学知识、信息，借助强大的数学工具和连贯的理论对实质性的物理知识、信息加以综合、融合，进行时空分析。鉴于某些特定的甲状腺癌症是由环境暴露引起的，通过暴露风险和甲状腺癌症时空分布的随机协同状况来建立标准，评判环境暴露风险的显著性，并根据环境暴露风险和甲状腺癌症的强度相关联来评判、预测健康风险。本研究以杭州市甲状腺癌症发病率和各种环境暴露风险为案例，证明其在评估环境暴露风险关联性上随机暴露分析中的价值，揭示物理系统的时空模型与人类健康风险分布之间的关联性，增进对“暴露-健康风险”的定量化理解，丰富空间流行病学理论，为甲状腺肿瘤等非传染性疾病的空间流行病学提供方法学支持。

近几十年来，甲状腺癌发病率在世界范围内持续上升。2008年-2012年中国浙江省杭州市共有有7147甲状腺癌病例数。因而，分析甲状腺癌的时空分布特征并探索相关的环境暴露因子具有重要的公共卫生意义。相关分析、方差分析、泊松回归被用来评估甲状腺癌发病率与主要环境暴露因子的统计学联系；贝叶斯最大熵方法被用来评估“暴露-患病”关联度的空间分布强弱；地理加权回归方法也被用来探索与甲状腺癌发病率的环境暴露因子。获得以下结果：高甲状腺癌发病率聚集区分布在杭州东北部、城市区域，而低发病率聚集区分布在杭州西南部、农村区域；经济社会指标与甲状腺癌发病率有显著正相关关系（r=0.687, p<0.01）；相比于低社会经济指标区域，中高指标区域具有较高的发病率，即相对风险为2.29（95%CI为1.99~2.63）、3.67（95%CI为3.22~4.19）；工业密度与甲状腺癌发病率的关系不显著；利用贝叶斯最大熵方法得到的预测误差结果表明引入社会经济指标能获得更低的发病率预测误差，表明社会经济指标与甲状腺癌发病率具有较强的关联性。通过地理加权方法分析得到工业密度、化学需氧量、建筑面积占比对杭州东北部郊外具有强的正向影响，而高程、坡度、森林面积占比在杭州中部和农村地区具有显著的负相关关系。弄清对甲状腺癌发病率影响较大的环境暴露因子有助于公共卫生部门决策制定。本项目研发的方法可以丰富空间流行病学理论，为甲状腺肿瘤等非传染性疾病的空间流行病学提供方法学支持。