Characteristics And Predictive Model Of Pulmonary Solitary Pure Ground-glass Nodules

Purpose:At present,different guidelines for the treatment of isolated pure ground glass nodules with the maximum diameter ≤3cm vary greatly,and there is no unified conclusion on the surgical timing of lung pGGN,which mainly depends on the experience of clinicians.In this paper,the clinical and CT characteristics of single pGGN were retrospectively analyzed to conclude whether the pathology of pGGN is a mathematical model of pre-invasive lesions,so as to provide a reliable basis for clinicians to diagnose and treat pGGN.Methods and materials:Retrospective analysis was performed on the clinical data and imaging features of 98 patients with isolated pure pulmonary ground glass nodules(SpGGN)admitted to the department of thoracic surgery of the second hospital of jilin university from January 2019 to December 2019.All patients underwent minimally invasive thoracoscopic surgery,and the pathological diagnosis was clear.According to the pathological types,they were divided into pre-invasive lesions(benign /AHH group and adenocarcinoma in situ /AIS group),invasive lesions(micro-invasive adenocarcinoma /MIA group and invasive adenocarcinoma /IACgroup).STATA15.0 software was used for statistical analysis.Age,maximum GGN diameter,gender,symptoms,smoking history,family history,pulmonary lobe where GGN is located,shape of GGN,pleural traction sign,smooth GGN edge,foliation sign,burr sign,clear or fuzzy boundary,cavitation sign,vascular cluster sign,and bronchopneumatic sign.The willcoxon rank-sum test was used for the comparison between the measurement data groups,and the chi-square test or Fisher’s exact probability method was used for the comparison between the classification variables groups.P < 0.05 was considered statistically significant.The single factor analysis was statistically significant and variables judged by clinical experience were included in the multi-factor analysis.The stepwise method was used to screen variables and explore the model.VIF,goodness of fit and AIC were used for model diagnosis and comparison,and the optimal model was obtained.ROC curve was used to evaluate the effectiveness of the model.Results:In this study,a total of 98 patients with isolated,pure ground glass nodules of the lung were collected,including 1 case of chronic inflammation and 22 cases of adenocarcinoma in situ.There were 27 cases of microinvasive adenocarcinoma and 48 cases of invasive adenocarcinoma.(1)univariate analysis: general clinical data: there was no statistical significance in gender,smoking history and family history oftumor between the infiltrating prelesion group and the micro-infiltrating adenocarcinoma/infiltrating adenocarcinoma group(P > 0.05),but there was statistical significance in age(P < 0.05).There were statistical differences in tumor markers(P=0.03).(3)imaging features: the maximum diameter,shape,pleural traction sign,burr sign,lobulation sign,and cavitation sign of GGN were statistically different in the preinvasive lesion group and the micro-invasive adenocarcinoma/invasive adenocarcinoma group(P < 0.05).However,there was no statistical significance in pGGN boundary,vascular cluster signs and bronchopneumatic signs(P > 0.05).(2)multiple factors analysis: gender,GGN shape,boundary,maximum diameter,burr sign,pleural lesion before force prediction is to distinguish the infiltration and tiny infiltrating adenocarcinoma/infiltrates the gonads in cancer independent risk factors,which in addition to the gender is to protect the factors,the rest are predicting small infiltrating adenocarcinoma/infiltrates the gonads in cancer risk factors.(3)establishment of a prediction model of lesions before and after infiltration:minimal infiltration/infiltration =-2.038491+0.1536551* maximum diameter +1.910714* burr sign +1.715138* pleural traction sign +1.12301*shape +1.129196* boundary-1.100775* sex.When the prediction probability was less than 0.691,pGGN was considered as a pre-invasive lesion,the area under the curve was 0.871,the sensitivity was 84.21%,the specificity was 86.36%,the positive likelihood ratio was 6.18,the negativelikelihood ratio was 0.183,and the jorden index was 0.705.Conclusion:1.There was no significant statistical difference in symptoms,smoking history,family history,vascular cluster signs,and bronchopneumatic signs between the two groups before and after the invasion of solitary pure ground-glass nodules.2.Cytokeratin fragment 19,age,foliation and vacuolation were statistically significant in determining whether SpGGN was MIA/IAC.3.Gender,SpGGN shape,boundary,maximum diameter,burrs,and pleural traction were independent risk factors for predicting the differentiation of pre-invasive lesions from micro-invasive adenocarcinoma/invasive adenocarcinoma.4.The prediction model of minimal infiltration/infiltration=-2.038491+0.1536551* maximum diameter +1.910714* burr sign+1.715138* pleural traction sign +1.12301* shape +1.129196* boundary-1.100775* gender had a good prediction effect on MIA/IAC.When the prediction probability ≥0.691,the lung SpGGN was considered to have a minimal invasive adenocarcinoma/invasive adenocarcinoma.When the prediction probability was < 0.691,lung SpGGN was considered as a pre-invasive lesion.