CT Features Combined With Tumor Markers For Predicting Invasiveness In Lung Cancers Of Ground-glass Opacity Nodules

Objective To explore the relevant risk factors that might predict tumor invasion of pulmonary ground glass nodules（GGO）among multiple factors and to establish Logistic regression model based on influencing factors of tumor invasion of pulmonary ground glass nodules（GGO）,so as to provide patients who is accidentally screened for small pulmonary nodules with the risk assessment of tumor invasion and assist in the diagnosis and treatment of small pulmonary nodules.Methods We collected and retrospectively analyzed the related data of 512 patients with lung nodules admitted to department of Thoracic Surgery in the First Hospital of Lanzhou University from 2018 to 2021 with definite pathological findings.Research data includ clinical data,imaging features（Chest CT data）and tumor markers.A total of 242 patients were eligible for inclusion in this study after screening on the basis of inclusion and exclusion criteria,including 107 males and135 females,with an average age of 57.98±9.571.The patients were divided into two groups,preinvasive lesions: adenomatoid hyperplasia（AAH）/adenocarcinoma in situ（AIS）and invasive lesions: microinvasive adenocarcinoma（MIA）/adenocarcinoma（IAC）.Firstly,univariate Logistic regression analysis was applied,and the elements with P <0.1 were further analyzed by multivariate Logistic regression analysis to further screen for the factors related to tumor invasion,and the prediction model was established using the screened factors.ROC curves of single predictive factors and predicted values of the model are drawn respectively and calculate the critical value.The sensitivity and specificity corresponding to the maximum value of the Yoden index were the sensitivity and specificity of the prediction model.Results Among 242 nodules included in this study,68 were preinvasive lesions,174 were infiltrating disease.There were 61 cases of pure ground glass nodules（p GGNs）,including 50 cases of pre-invasive lesions（16.6%）and 21 cases of invasive lesions（8.7%）.There were 181 cases of partial solid nodules（m GGNs）,including 28 cases of preinvasive lesions（11.6%）and 153 cases of invasive lesions（63.2%）.The results of univariate analysis showed that the proportion of infiltrating lesions in partial solid nodules（m GGNs）was significantly higher than that in pure ground glass nodules（p GGNs）.Age,standard diameter,maximum diameter of solid component,volume,smoking history,lobulation,and Cy FRA21-1 were also associated with tumor invasion.Mean CT value,maximum CT value,central CT value,and burr were highly correlated with tumor invasion,with P <0.001.There were no significant differences between the two groups in gender（P=0.009）,NSE（P=0.58）,Pro-GRF（P=0.242）,CEA（P=0.194）,pleural traction（P=0.235）,and vessel passage（P=0.137）.Multivariate Logistic regression analysis showed that independent risk factors for the diagnosis of tumor invasiveness ground-glass opacity nodules were Central CT value,Cyfr21-1,maximum diameter of solid component,nodule nature and burr of the nodules.The Optimum critical value of the above indicators between AAH/AIS and MIA/IAC were-309.0HU,3.23ng/m L,8.65 mm.The prediction model formula for tumor invasiveness probability was logit（P）=0.982-（3.369 x nodule nature）+（0.921 x maximum diameter of solid component）+（0.002 x Central CT value）+（0.526 x Cyfra21-1）-（0.0953 x burr sign）.The areas under the curve（AUC）obtained by plotting the receiver operating characteristic curve（ROC）using the regression probabilities of regression model was 0.908.The accuracy rate was91.3%,Sensitivity,specificity and accuracy were 91.9%,89.7% and 91.3%respectively.Conclusion The prediction model of lung ground glass nodular tumor invasion established in this study has high reliability,among which the nodular nature,central CT value,Cy FRA21-1,burr sign and maximum diameter of solid component have great clinical application value for the diagnosis of lung nodular tumor invasion.Partial solid pulmonary nodules with maximum diameter of solid component≥8.65 mm,central CT value ≥-309.0HU and marginal burr are more likely to be invasive lung cancer and surgical treatment should be performed as soon as possible in the absence of obvious surgical contraindications.otherwise CT follow-up observation.The logistic regression model established in this study can better predict the tumor Invasiveness of GGOs lung adenocarcinoma on CT and tumor markers.the predictive value is significantly higher than the independent use of each quantitative factor...