Spatiotemporal Correlation Analysis Of Environmental Factors And Thyroid Cancer Risk In Hangzhou City

Background:The occurrence of cancer is closely related to the pollution of natural environment and the circumstance of social environment.Thyroid cancer(TC)as the most common endocrine malignancy tumors,in the past three decades,its incidence showed a fast increasing rate,which has been a serious public health issue.Though numerious studies were conducted to explore TC pathology,seldom study was done in the aspects of spatiotemporal distribution and soil-geography environment.As the complexity of cancer etiology,new methods and techniques are demanded to analyze cancer and risk data more sufficiently.As the development of geographic information system(GIS)and its strong intergration ability,which could combine medical,epidemiology,geography,mathematics and computer science,making it possible for researchers to detect the spatiotemporal distribution of cancer,relate to corresponding environmental risk factors and provide foundations for cancer etiology exploration,cancer prevention and public health resources allocation.Methods:Hangzhou was selected as the study area,including 8 districts and 5 counties.There are totally 200 townships administrative unit(sud-district,town and country)in the study area and the resident population is about 8.7 million.All new TC cases[International Classification of Disease 10th edition(ICD-10)code:C73]diagnosed in Hangzhou during 2008-2012 were obtained from the Center for Disease Control and Prevention(CDC),Hangzhou city.After calculating the age-standardized incidence,Chi-square test and Poisson regression were employed to compare incidence disparities among various sub-populations and regions.After that,Moran’s I,Hotspot Getis-Ord Gi*,Anselin Local Moran’s I,Space,Space-time and space variation Scan Statistics were implemented to detect the spatiotemporal distribution pattern of TC incidence.Geo-Detector was employed to assess the associations of topographic and soil material factors to TC risk.Finally,combined with socioeconomic status(SES),industry density(ID)and female breast cancer incidence(BC)data,Spatiotemporal Random Field(S/TRF),Bayesian Maximum Entropy(BME)and Integrative Disease Predictability(IDP)were used to quantitatively assess the effects of risk factors on TC incidence.Results:TC incidence increased fast in Hangzhou,during the period 2008-2012,with the overall annual percentage change of 34.47.Female TC incidence was significantly higher than males,with female to male ratio about 3.2.People living in subdistrict area showed significantly higher TC incidence,was about 1.8 times of that in town area and about 3.1 times of that in country area.Population living in normal iodine background area showed significantly higher TC incidence,was about 2.7 times of that in moderate iodine deficient area and about 2.2 times of that in severe iodine deficiency area.TC revealed a prominent spatiotemporal distribution pattern.Downtown areas were high incidence cluster areas with significantly higher TC risk,whereas large parts of southwestern areas were low incidence cluster areas with significantly lower TC risk.Besides,Linping subdistrict was a spatial outline where showed high TC incidence itself but enclosed by low TC incidence townships("High-low"),the northern suburb areas of downtown although with low TC incidence,showed significantly higher increasing rate than other areas.Elevation,slope,soil parent material,organic carbon and total nitrogen are all significantly associated with TC distribution.TC incidence decreased as elevation and slope increased.Concerning soil parent materials,deposited materials were generally linked to higher TC incidence.SES had a significantly positive correlation with TC risk,Compared to people living in low SES areas,those living in median SES areas(relative risk(RR):2.29,95%confidence interval(95%CI):1.99-2.63)and high SES areas(RR:3.67,95%CI 3.22-4.19)showed significantly higher TC risk.According to the results of IDP,the positive correlation between SES and TC existed at both global and local scales,moreover,the strongest SES and TC correlation were distributed among residents mainly on a stripe of land from northeast to southwest(consisting mainly of sub-district areas).ID and TC correlation was not detected,the results of IDP didn’t support the correlation between ID and TC neither at global nor at local scales.Finally higher spatiotemporal correlation was found between female TC and BC,The results of IDP showed that population with higher female BC incidence related to higher female TC incidence and vice versa.Conclusions:The fast increasing of TC incidence which could be attributed to both over-diagnose and true environmental risk factors,should gain sufficient attentions from scientists and governments.More studies should be conducted to explore the etiologies of TC.TC incidence showed big disparities among different areas and sub-populations in Hangzhou,more attentions should be paid to high risk populations,to Linping subdistrict,and to the areas of northern suburb areas of downtown where the highest increasing rate was detected.SES usually as a proxy for health care accessibility,at least partly explained that over-diagnose was responsible for high TC incidence,especially on the strip of land where stronger SES-TC correlation was detected.The present study lacks of information about types and amounts of the chemicals that those industry factories discharged.Further studies should focus on specific chemicals and evaluated the association between thyroid cancer risk and detail chemicals exposures.Higher BC incidence which could be seen as proxies for hormone and reproductive risk factors,were positive correlated with TC incidence,thus hormone level and reproductive factors might be responsible for TC risk,when some area showed higher TC/BC incidence,more attention should be paid to another(BC/TC)cancer.