Epidemiological Analysis Of 3317 Cases Of Pulmonary Nodules And Validation Of The Risk Prediction Models Of Lung Cancer

Objectives:Early screening can help reduce the mortality rate of lung cancer,for which computed tomography(CT)is the main method.The objectives of this study were to explore the epidemiological characteristics of pulmonary nodules found by lung CT in the physical examination of healthy population,and validate the application value of three different lung cancer risk prediction models.Methods:We retrospectively analyzed the clinical data of healthy population who underwent lung CT examination from August 2015 to July 2016 in the physical examination center of Chinese PL A General Hospital.The population was divided into pulmonary nodule group and non-pulmonary nodule group on the basis of imaging findings.Data collected for epidemiological investigation included gender,age,smoking history,disease history,family history of tumor,and the number and location of pulmonary nodules,of which the statistical analysis was performed with SPSS 24.0 and STATA/SE 12.0.The classical Mayo Clinic model,VA model and Peking University model formula in China were validated with clinical data of patients bearing lung nodules detected by physical examination and confirmed by pathological diagnosis of the hospital.The sensitivity,specificity,positive predictive value and negative predictive value of three different lung cancer risk prediction models were compared.SPSS 24.0 was used for statistical analysis.Results:A total of 19637 cases underwent lung CT examinations in the physical examination center of our hospital,which were classified into pulmonary nodule group(3317 cases,16.9%)and non-pulmonary nodule group(16320 cases,83.1%).The most patients(1394 cases,42.0%)were aged between 50 and 59 years in the pulmonary nodule group while the most people(6017 cases,36.9%)were aged between 40 and 49 in the non-pulmonary nodule group.The mean age of pulmonary nodule group was significantly older than that of non-pulmonary nodule group(51.1 vs 46.8 years).Patients with tumor history and pulmonary diseases accounted for 2%(49 cases)and 3%(91 cases)of the pulmonary nodule group,respectively,while those with tumor history and pulmonary diseases accounted for 1%(155 cases)and 1%(141 cases),respectively.There were significant differences in disease composition and mean age between pulmonary nodule group and non-pulmonary nodule group(P<0.05).The proportion of patients who had disease history was higher in pulmonary nodule group than that in non-pulmonary nodule group while there was no significant difference in gender,smoking history and family history of tumor between the two groups(P>0.05).In the pulmonary nodule group,there were significant differences in disease composition and mean age between patients with single pulmonary nodule and those with multiple pulmonary nodules(P<0.05).Patients with disease history were more likely to have multiple pulmonary nodules compared with single pulmonary nodule(4.2%vs 2.0%).Patients with multiple pulmonary nodules were significantly older than those with single pulmonary nodule(52.0 vs 50.7 years)while there was no significant difference in gender,smoking history and family history of tumor(P>0.05).The mean age differed significantly among patients with different sites of pulmonary nodules(P<0.05).The risk factors for pulmonary nodules included advanced age,smoking history and disease history.Three lung cancer risk prediction models were validated by 407 patients whose pulmonary nodules were found by physical examination and confirmed pathologically.The sensitivity of Peking University model was the highest(78.6%)while the specificity of Mayo Clinic model was the highest(76.2%).The area under ROC curve of Mayo Clinic model,VA model and Peking University model was 0.608,0.550 and 0.615,respectively.Conclusion:Compared with non-pulmonary nodule group,the incidence of pulmonary nodules is related to the mean age.The incidence of single pulmonary nodule or multiple pulmonary nodules is related to disease history and mean age.Pulmonary nodules are prone to occur in the elderly.Patients who have multiple pulmonary nodules involving both lungs are significantly older than those with single pulmonary nodule.Patients who have pulmonary nodules involving the right lung alone are significantly older than those with pulmonary nodules involving the left lung alone.Advanced age,smoking history and disease history are risk factors for pulmonary nodules.Among the three models validated in the study,Peking University model has the highest sensitivity while Mayo Clinic model has the highest specificity,and the diagnostic value of Peking University model is better than that of VA model.For pulmonary nodules found in the physical examination of healthy population,evaluation of malignant probability should be considered comprehensively.