The Analysis Of Risk Factors And The Construction Of Clinical Predictive Model For Pulmonary Nodules

Objective:To find the risk factors of malignant pulmonary nodule and establish the predictive model by analyzing the general clinical information and radiography of patients who had pulmonary nodules,providing diagnosis reference of pulmonary nodule for clinician.Methods:A retrospective study was conducted in 273 patients who got definite pathological diagnosis by surgical operation of pulmonary nodules from January 2019 to December 2019 formed group A.They were separated into two groups called the benign group(n=66)and the malignant group(n=207)according to pathological diagnosis.The following factors were evaluated:age,gender,history and quantity of smoking,history of pulmonary disease,family history of lung cancer,family history of other tumor and radiography data including morphology,border,density,location,quantity of local nodule,diameter,special imaging sign(lobulation,burr,pleural indentation,bronchotomy sign,vascular cluster sign,calcification and vacuole sign),lymphadenectasis of mediastinal or porta pulmonis,enhanced CT scan radiography.Single factor analysis and multivariate logistic regression analysis were used to find the risk factors affecting the definite pathological diagnosis of solitary pulmonary nodules.In addition,100 patients with definite pathological pulmonary nodules from surgical resection at our hospital in 2020 were randomly selected as group B to verify the clinical predictive value of this model.Results:In the analysis of single factor,there were significant differences in age,sex,smoke,morphology,border,density,diameter,lobulation,burr,pleural indentation,vascular cluster sign,vacuole sign and bronchotomy sign between the patients with benign and malignant pulmonary nodules(P<0.05).Logistic regression analysis showed that age,diameter,lobulation,burr,pleural indentation and nodule density were independent risk factors for malignant pulmonary nodules,and its predictive model could be established as follows:P=ex/(l+ex),x=-4.323+(0.042× age)+(0.945 X diameter)+(0.963×lobulation)+(1.035×burr)+(1.163×pleural indentation)+(2.230 X nodule density).The data of group B were substituted into the prediction model of this study,Li Yun model and Mayo model,according to which the ROC curve was drawn.The AUC of the prediction model of this study was 0.908±0.043,and the 95%confidence interval was(0.823-0.994).The sensitivity,specificity,positive predictive value and negative predictive value of this model were 91.7%,81.2%,96.3%and 65.0%,respectively.The AUC of Li Yun model is 0.698±0.066,while Mayo model is 0.765±0.051,so the predictive model of this study is more accurate than Li Yun model and Mayo model.Conclusions:Age,diameter,lobulation,burr,pleural indentation and nodule density are independent risk factors for malignant pulmonary nodules.And the predictive model of malignant pulmonary nodules has a better sensitivity and specificity,which has a great significance for the judge of benign and malignant pulmonary nodules.