| Numbers of researches in the field of Environmental Medicine showed that both historic lung disease and genetic factors,and environmental factors including smoking,occupational exposure,as well as environmental pollution such as atmospheric pollution are the important contributor factors to lung cancer.Smoking has been demonstrated to be a main contributor to lung cancer,however which can be deducted for its contributions to lung cancer due to the actions of Smoking Control and Non-smoking in China.During the past decades,the slowly varying and declining smoking rate in China is not consistent with significant increasing lung cancer morbidity.The historical lung disease and genetic factors seems not associated significantly with and lung cancer morbidity on a national scale as well.So the effect of these factors on lung cancer morbidity has been ignored in this study.Increasing air pollution levels associated with rapid urbanization and industrialization in China have severely deteriorated the urban environment across the country,causing increasing lung cancer risks of the Chinese population,especially female lung cancer morbidity.Among the atmospheric environmental factors,fine particulate matter(PM2.5)appeared to be a leading factor to growing lung cancer morbidity.In 2013,the International Agency for Research on Cancer（IARC）officially classified PM2.5 as a first-class carcinogen（first-class definition:identified carcinogen for human beings）.In addition to affecting cardiovascular and respiratory diseases,PM2.5 also has exhibited a close relationship with the female lung cancer morbidity.It is now widely recognized that long-term exposure to PM2.5 could lead to increased risk of lung cancer morbidity.In the present thesis study,we mainly focused on the impact of the annual average PM2.5 concentration on the temporal and spatial distribution of female lung cancer morbidity based on the satellite retrieved PM2.5（longitudinal and latitudinal resolution0.01°×0.01°）and the statistical data of female lung cancer morbidity from 2010 to2014 in China,and the temporal and spatial relationship between historical PM2.5exposure concentration and female lung cancer morbidity were established as well.Firstly,the correlation between PM2.5 and the female lung cancer morbidity was studied by association analysis methods.A spatial autocorrelation method was used to evaluate the spatial relationship between the female lung cancer morbidities from 2010to 2014 and satellite-derived ground-level of PM2.5.Annual mean PM2.5 levels in this specific year and previous 8 years were selected as 9 independent variables（Lag0,Lag1,Lag2,Lag3,Lag4,Lag5,Lag6,Lag7,Lag8）to study the female lung cancer morbidity based on correlation analysis and grey comprehensive correlation analysis.For further research of the impact of PM2.5 on the female lung cancer morbidity,the fuzzy mathematics method and maximum entropy model were applied to identify the importance of influencing factors for the female lung cancer morbidity across China from 2010 to 2014.The fuzzy membership functions between the female lung cancer morbidity and 9 influencing factors（Lag0,Lag1,Lag2,Lag3,Lag4,Lag5,Lag6,Lag7,Lag8）were generated and the membership degree of each factor was calculated as well.Based on the maximum entropy model,the weights of 9 influencing factors from 2010 to 2014 were established respectively to confirm the importance of each factor.In order to evaluate the effect of 9 influencing factors comprehensively,the environmental suitability index of PM2.5 for female lung cancer morbidity was calculated by weighted average method.We categorized the environmental suitability combined with Natural Breaks analysis into three zones,including low-suitable region,medium-suitable region,as well as high-suitable region where the female lung cancer mortality ranging from low to high rate was identified.Based on the female lung cancer morbidity from 2010 to 2014,we established ridge regression,support vector regression and BP artificial neural network model to predict the female lung cancer morbidity in China for 2015 and 2016.And standard deviation method,variance countdown method and optimal weighting method were respectively used to weigh the three methods in order to build the best combined predicton model.To evaluate model performance and choose the optimal model,the error analysis has been carried out to measure the difference between the measured and predicted female lung cancer morbidity.Finally,the female lung cancer morbidities of2051 cities in the research area from 2015 to 2016 were predicted by using the optimal model,and then generate the spatial distribution trend of morbidity.The results are as follows:（1）The association analysis showed that the spatial difference of female lung cancer morbidity was obvious which was not distributed randomly in study area.The correlation between PM2.5 levels and female lung cancer morbidity in the period of lagging third year and sixth year（lag3-lag6）was as strong as those estimated lag correlations,especially the lag6,the grey comprehensive correlation coefficient which also indicated that in the period of lagging sixth year and eighth year（lag6-lag8）had strong influence on the female lung cancer morbidity.The overall results of association analysis preliminarily proved that lag6 had a strong correlation with the female lung cancer morbidity.（2）The importance of influencing facors analysis showed that the membership degrees of lag0,lag1,lag2,lag3,lag4,lag5,lag6,lag7 and lag8 were different after standardization by the membership functions,but the trends were similar.The high-value areas of membership degree were mainly concentrated in Tianjin city and southeast of Hebei province and its surrounding areas,it indicated that there existed a PM2.5 environment which was most suitable for inducing lung cancer in these areas.The maximum information entropy model was used to calculate the weight of 9influencing factors.The results showed that lag6 with the largest weight ranked top among the five-year weight calculation,which indicated lag6 was the most significant factor among the 9 factors.（3）The results of environmental suitability assessment and analysis showed that the environmental suitability index of PM2.5 for female lung cancer morbidity was0.2069-0.9329.The study area was divided into low-suitable,medium-suitable and high-suitable zones according to environmental suitability.The high-suitable area,mainly distributing in the Tianjin city-Hebei province and North China Plain which was the heaviest contaminated region by air pollution in China,covering most of the provincial capitals and cities in the research area,especially the Tianjin-Hebei region,accounting for 39.44%of the total area investigated in this thesis study.Correspondingly,35.22%of the total area was medium-suitable zone and 25.34%was low-suitable zone.（4）The evaluation of prediction model and spatial estimation suggested that the variance countdown method of combined prediction model performed as the best method in the female lung cancer morbidity forecast.We calculated the female lung cancer morbidity by the model among the 2051 sites of research region.The morbidity distributions from 2015 to 2016 across the research region were obtained by spatial valuation method.The results showed that the predicted female lung cancer morbidity in the study area was 10.2634 to 41.5679 for 2015 and from 21.0599 to 50.9948 for2016.The overall trend of female lung cancer morbidity was increased significantly from western to eastern,and the high-value areas were concentrated in southern Hebei,Shandong,northern Jiangsu,Anhui provinces,and Eastern Henan province.Generally,the cumulative effect and lagging effect of PM2.5 on female lung cancer morbidity were obvious.With the increase of lagging years,the temporal-impact trend of PM2.5 increased from weak to strong,and then decreased gradually after reaching its peak.The spatial trend distribution of female lung cancer morbidity was consistent with the results of suitable zoning.The high morbidity area was mainly concentrated on the high-suitable zones with high PM2.5 concentration.At the same time,it was also the heavily polluted zones and densely populated areas.It can be seen that PM2.5 posed a significant temporal and spatial effect on the female lung cancer morbidity.This thesis study of temporal-spatial distribution of PM2.5 concentration for female lung cancer morbidity could provide a scientific basis for the environmental management,health risk assessment,and female lung cancer control. |
