Study On Risk Factors Of Lung Cancer And Construction Of Risk Prediction Model Based On NHANES And Meta-analysis

Background:Lung cancer is the malignant tumor with the highest mortality rate in the world,and its incidence rate remains the highest in China.The occurrence of lung cancer is the result of multiple factors,the etiology is complicated and the mechanism has not been clarified.Although the treatment of lung cancer has been significantly improved in recent years with the emergence of molecular targeted drugs and gene therapy,but surgery,radiotherapy and chemotherapy are still the most commonly used and basic treatment methods for lung cancer in China,and the prognosis is still poor,with the five-years survival rate remains below 17%.Foreign randomized controlled trials have shown that low-dose CT screening for lung cancer in high-risk groups can reduce lung cancer mortality by at least 20%.This shows the importance of primary prevention of lung cancer.With the progress and development of society,people’s living environment and habits are also changing constantly.It is of great significance to continuously explore and timely update the risk factors of lung cancer and apply them to the screening of high-risk groups,in order to prevent and control the occurrence and development of lung cancer.Objective:1.The purpose of mining foreign public databases is to discover new demographic and nutrition-related risk factors of lung cancer,so as to provides reference for the prediction model to choose which risk factors to include and it’s also provide inspiration and direction for lung cancer epidemiological research of China.2.On the basis of obtaining the risk factors of lung cancer in the Chinese population through systematic review and meta-analysis,a risk prediction model was constructed and verified externally,aiming at helping the screening of high-risk groups of lung cancer in China in the future and achieving the purpose of preventing lung cancer.Methods:1.Based on the National Health and Nutrition Examination Survey database of the United States,the demographic data,physical measurement data,biochemical indexes and dietary nutrition data of lung cancer patients and healthy people were mined by R 4.2.1software.Then lung cancer patients and healthy people were matched according to age,sex and race in a ratio of 1:5,and the data were analyzed using case-control studies to obtain the odds ratio and 95%confidence interval of risk factors related to lung cancer incidence in the US population.2.After setting search conditions,case-control studies of lung cancer risk factors of the Chinese population published from January 1,2010 to January 1,2022 were retrieved from well-known Chinese and foreign databases.After literature screening at all levels,the OR and 95%CI of lung cancer risk factors of each study were extracted.Review Manager5.3 and Stata 16.0 were used for meta-analysis.3.The OR value and 95%CI of risk factors obtained from the above meta-analysis were used to construct a logistic regression model,and the questionnaire data collected in the hospital were used to conduct external verification of the model,and the Receiver Operating Characteristic curve of subjects was drawn to evaluate the prediction ability of the model.Results:1.Multivariate logistic regression was performed on the data extracted from NHANES database.After adjusting gender,age and race,statistically significant results of each factor were obtained as follows:Previous smoking（OR=6.76,95%CI:3.10-14.73）,history of COPD（OR=2.77,95%CI:1.11-6.91）,ABSI between 0.087 and 0.132（OR=2.07,95%CI:1.03-4.14）were risk factors for lung cancer.BMI values between 25 and30（OR=0.33,95%CI:0.17-0.63）,ratio of family income to poverty>3.5（OR=0.38,95%CI:0.15-0.96）,daily total dietary calories between 2081.33-5122kcal（OR=0.30,95%CI:0.15-0.63）,total vegetable score between 2.509-5（OR=0.48,95%CI:0.24-0.96）and total vegetable scores equal to 5（OR=0.49,95%CI:0.25-0.96）were protective factors for lung cancer.2.Finally,34 literatures were included in the meta-analysis,and the OR values and95%CI of the included studies were combined.After adjusting for heterogeneity,the results were as follows:Smoking OR=3.29（95%CI:3.02-3.57）,passive smoking,OR=2.31（95%CI:2.03-2.62）,family history of cancer OR=4.36（95%CI:3.30-5.77）,respiratory disease history OR=2.89（95%CI:2.24-3.73）,depressed OR=2.7 8（95%CI:2.25-3.44）,OR=3.45（95%CI:2.73-4.35）,OR=1.87（95%CI:1.55-2.26）were the risk factors of lung cancer in Chinese population.Regular physical exercise OR=0.41（95%CI:0.34-0.48）,eating vegetables and fruits OR=0.41（95%CI:0.35-0.48）,drinking tea OR=0.66（95%CI:0.60-0.73）are protective factors for lung cancer in Chinese population.3.Based on the results of meta-analysis,the prediction model of lung cancer risk in Chinese population was established as follows:=⁰0)（）)₁₊₀0)（）),（）=-7.39+1.19₁+0.84₂+1.47₃-0.89₄-0.89₅-0.42₆+1.06₇+1.02₈+1.24₉+0.63₁₀(₁:smoking,₂:passive smoking,₃:family history of cancer,₄:regular physical exercise,₅:regular eating of fruits and vegetables,₆:drinking tea,₇:history of respiratory diseases,₈:emotional depression,₉:exposure to cooking fume,₁₀:nearby polluting factories).The area under ROC for external verification of the model was 0.859,the maximum Youden Index was 0.566,the corresponding sensitivity was 0.969,and the specificity was 0.403.Conclusion:1.The analysis of NHANES database innovativeness found the relationship between some new indexes such as ABSI,HEI,DII and the risk of lung cancer in the American population,which provided ideas for the future research on the risk factors of lung cancer in China.2.Meta-analysis results showed that smoking,passive smoking,family history of cancer,history of respiratory diseases,emotional depression,exposure to cooking fume,and nearby polluting factories were risk factors for lung cancer in the Chinese population.Regular physical exercise,eating vegetables and fruits,drinking tea are protective factors for lung cancer in the Chinese population.There are significant differences in lung cancer risk factors between Chinese and American populations.3.The prediction model of lung cancer risk in the Chinese population established by using the 10 combined OR value in meta-analysis has certain predictive ability,with an AUC of 0.859,a sensitivity of 0.969,and a specificity of 0.403.