| Objective: Univariate and multifactorial analyses were performed on the general clinical data,chest CT imaging features and tumour markers of incidental small pulmonary nodules.In order to develop a predictive model for the probability of malignancy of incidental small pulmonary nodules,a binary logistic regression-based approach was used in this study,and its efficacy was also tested and compared with the Mayo and Brock models.Methods: Reviewing the pathological distribution of 358 incidental small pulmonary nodules in the First Affiliated Hospital of Kunming Medical University after surgery from October 2020 to June 2022,this study collected clinical general information on patients,including age,gender,smoking history,family history of tumor and vocational exposure;chest CT imaging features of small pulmonary nodules: diameter of nodules,anatomical location of nodules,nodule type,spiculation,lobulation,vascular clustering sign,pleural indentation,vacuole sign,vacavity sign;tumour markers: NSE,CEA,CYFRA21-1.249 individuals in the training group with265 nodules and 65 individuals in the validation group with 83 nodules,using 31 December 2021 as the cut-off point.The variables that were compared between groups and identified as significant differences in the training group were considered as risk factors.In this study,the risk factors collected were subjected to binary logistic regression analysis to develop a predictive model for the probability of malignancy of small pulmonary nodules,which was validated and assessed by evaluating its area under the ROC curve and the Hosmer-Lemeshow test fit.The proposed model,Mayo model and Brock model were tested in the validation group,and their ROC curves were plotted to specify their respective AUCs,best cut-off values,as well as sensitivities,specificities,positive predictive values,negative predictive values,positive likelihood ratios and negative likelihood ratios.Results:Age,history of smoking,increase in nodule diameter,spiculation,vascular clustering sign,vacuole sign,ground glass nodules and subsolid nodules are all risk factors for malignant tumor and AIS.Modeling: P=ex/(1+ex)where x=-6.237+0.046×age+1.013×smoking history+0.266×nodule diameter(mm)+1.329×burr sign+1.943×vessel cluster sign+2.549×vacuolar sign+3.380×ground glass nodule+3.097×subsolid nodule.In the training group,the AUC was 0.917,the Youden index was 0.712,the diagnostic threshold was p=0.679,the sensitivity was 0.891,the specificity was 0.733,the positive predictive value was 0.867,the negative predictive value was 0.776,the positive likelihood ratio was 3.337 and the negative likelihood ratio was 0.149.In the test group,the AUC of the model was 0.900,the Youden index was 0.719,the sensitivity was 0.833,the specificity was0.886,the positive predictive value was 0.907,the negative predictive value was 0.80,the positive likelihood ratio was 7.307,the negative likelihood ratio was 0.1885;The Mayo model had an AUC of0.676,a Youden index of 0.277,a sensitivity of 0.563,a specificity of 0.714,a positive predictive value of 0.722,a negative predictive value of 0.703,a positive likelihood ratio of 1.969 and a negative likelihood ratio of 0.612.The AUC of the Brock model was 0.641,the Youden index was 0.324,the sensitivity was 0.896,the specificity was 0.429,the positive predictive value was 0.667,the negative predictive value was 0.750,the positive likelihood ratio was 1.569,and the negative likelihood ratio was 0.242.The model has good discrimination and calibration,and is significantly more effective than the Mayo and Brock models in the validation group.Conclusions: 1.Age,smoking history,spiculation,vascular collection sign,vacuolar sign,ground glass nodule and subsolid nodule are risk factors for incidental small pulmonary nodules;2.The Moya model and Brck model are not applicable to predict the benignity and malignancy of incidental small pulmonary nodules in China;3.The mathematical model constructed in this study,to a certain extent,assist clinicians in assessing the probability of malignancy of incidental small pulmonary nodules and reduce underdiagnosis,misdiagnosis and excessive surgical intervention. |
PDF Full Text Request