Establishment of a model based on MSCT three-dimensional image combined with logistic regression analysis for the clinical predictive value and prognosis of lung adenocarcinoma subtypes

CHEN Wu; MIN Xuhong; SONG Biao; WANG Bin; LI Yuan; SONG Qilong; WANG Zhaohua

Abstract：Objective To investigate the clinical predictive value and prognosis analysis of MSCT 3D image feature perfor-mance combined with logistic regression analysis and to establish a model in lung adenocarcinoma subtype.Methods The qualitative and quantitative data of clinical data and three-dimensional imaging of patients with pathologically confirmed lung in-vasive adenocarcinoma and the longest diameter of nodules≤3.0 cm in Anhui Provincial Chest Hospital were selected for com-parative analysis,and the risk of lung adenocarcinoma subtype was mainly divided into MP/S+group(Contains micropapillae components or solid ingredient groups)and MP/S-group(Contains no micropapillae or solid ingredient groups)for comparative analysis.The independent predictors were screened out by logistic regression analysis and the prediction model was established,and then the ROC curve was plotted to compare the efficacy of the two groups of IAC,and finally,the prognosis analysis was ob-tained according to whether there was invasion of neighboring tissues after surgery.Results Among the 206 patients with inva-sive adenocarcinoma,the longest diameter was≤3.0 cm,of which 89 patients were MP/S+group,117 patients were divided into MP/S-group,the average age of onset was(59.0±9.53)and(60.6±10.3)years,respectively,and there was no signifi-cant difference between the longest diameter of nodules and lobulation signs between the two subtypes(P>0.05).The difference between the proportion of solid components>50%and burr signs in the classification of the two subtypes was statistically significant(P<0.05).The results of logistic regression analysis show that>50%of the solid components were independent predic-tors,while the ROC curve showed that the predictive value of the logistic model was higher than that of the independent predic-tors.Conclusion The qualitative and quantitative data of MSCT 3D images combined with logistic can improve the diagnostic rate of MP/S+group,so as to predict prognosis and to improve survival.