Abstract: Objective To detect the correlation between the clinical imaging features and the degree of invasion of pulmonary pure ground-glass nodules (pGGNs) (≤ 3cm). Methods The  clinical data of 886 patients who underwent minimally invasive surgery from June 2013 to June 2016 for pulmonary nodules were collected. Among these, 72 patients had complete clinical data and isolated pulmonary GGN resection, and the  diameter of the pulmonary GGN was ≤ 3cm.Logistic regression analysis was used to analyze the independent risk factors for invasiveness of pure ground glass  nodules. Receiver operating characteristic curve (ROC curve) was used to determine the optimal cut-off value for preinvasive and invasive lesions. Results Univariate analysis demonstrated that there were significant differences in carcinoembryonic antigen,maximum diameter,and area of pGGNs in patients with pre-invasive and invasive lesions (P<0.05).Multivariate logistic regression analysis showed that there were only statistical differences in the maximum diameter of nodule pre-invasive and invasive lesions.The optimal cut-off value for CT-maximal diameter to predict pre-invasive lesions or invasive lesions was 1.08cm. Conclusion Therefore,predicting the pathological types of nodules (pre-invasive or invasive) by measuring the maximum diameter of pGGNs is reliable,with the most reliable cut-off value as 1.08cm.

Key words: lung cancer, high-resolution CT, pure ground-glass lung nodule, pathological diagnosis