| Objective: To establish a quantitative diagnosis model for the diagnosis of lung invasive adenocarcinoma with ground glass nodule(GGN).Methods: 278 cases of GGNs were analyzed retrospectively,including 186 cases of invasive lung adenocarcinoma(IAC)and 92 cases of non-invasive adenocarcinoma(including minor invasive adenocarcinoma MIA).The density,volume,mass,2D long and short diameter,3D axial diameter and standard diameter of the nodules were obtained by using the artificial intelligence recognition software.Mann Whitney U test was used to compare the parameters of IAC group and non-IAC group,and multivariate logic analysis was used to establish quantitative parameter model and plot ROC curve.Result: There were significant differences between the IAC group and non-IAC group for all the parameters(P < 0.001).After multiple logistic analysis,the density and 2D length were included to establish the quantitative diagnosis model.The ROC curve showed that the area under the curve(ROC)of the model was 0.819,which was higher than the diagnostic efficiency of each single parameter.Conclusion: The quantitative diagnosis model based on the density and 2D long diameter can effectively diagnose the invasive adenocarcinoma with ground glass density node. |