Risk Factors And Prediction Model Of Single Subsolid Pulmonary Nodules

Objective:The purpose of this study was to explore the clinical characteristics and related risk factors of patients with single subsolid pulmonary nodules during medical treatment.Then we developed a mathematical prediction model for the early diagnosis and treatment of single subsolid pulmonary nodules.Methods:Data from a total of 208 patients with benign or malignant single subsolid pulmonary nodules were retrospectively collected.The patient's data includes clinical symptoms(age,gender,smoking history,tumor history),radiological features(nodule location,nodule size,nodule boundary is clear,with or without glitches and lobes,and blood vessel penetration)and other 17imaging performance and laboratory indicators(CEA,NSE,Cyfra21-1,SCCA,CA199,CA125,CA153).Quantitative data were analyzed by independent sample t-test and non-parametric test to analyze the differences between groups;qualitative data were selected by chi-square test to analyze the differences between benign and malignant nodules.The method of multiple stepwise logistic regression analysis was used to analyze the statistically significant predictors from single factor analysis,and finally a mathematical prediction model was established.Results:There were 17 patients with benign nodules in 208 patients,while the remaining 191 patients were malignant nodules.Independent predictors for malignant subsolid nodules included NSE,nodule size,blood vessels in the nodules,mixed ground glass nodules and spicular sign.The predictive mathematical model was established:P=e~X/(1+e~X),X=-10.809+(2.955×mixed ground glass nodule)+(3.007×vascular nodule penetration in the nodule)+(3.828×minutes Ye Zheng)+(0.138×nodule maximum diameter)+(0.824×NSE),e is the natural constant.By plotting the ROC curve,the area under the curve was 0.974,and the 95%confidence interval was 0.951 to 0.997.The results showed that the model had better prediction ability.Conclusions:The independent risk factors for predicting malignant solid nodules included the size of single subsolid pulmonary nodule,blood vessels passing through the nodule,mixed ground-glass nodules,image performance of spicular sign and values of neuron specific enolase.The diagnostic accuracy of single SSN mathematical prediction model which separated benign subsolid nodules from the malignant was satisfactory.The result of this prediction model was based on multiple logistic regression.