Construction Of A Nomogram Model For Predicting The Invasiveness Of GGN Lung Adenocarcinoma Based On HRCT Features

Objective(s):a model for predicting invasiveness of ground-glass nodule(GGN)was constructed based on quantitative and qualitative features of Computed tomography(CT)imaging to provide reference value for preoperative planning of GGN patients.Methods:1.Clinical data of patients with GGN lung adenocarcinoma who underwent surgical treatment in the Third Affiliated Hospital of Kunming Medical University from September 1,2020 to November 30,2022 were collected and analyzed according to inclusion and exclusion criteria;Patients who had surgery from September 1,2020 to July 31,2022 were included in the modeling group,and patients who had surgery from August 1,2022 to November 30,2022 were included in the validation group.According to the difference of invasiveness,Adenocarcinoma in situ and microinvasive adenocarcinoma were classified as non-invasive adenocarcinoma group.Lepidic predominant,papillary predominant,acinar predominant,solid predominant and micropapillary predominant were classified into invasive adenocarcinoma group.2.Collection and analysis of quantitative and qualitative CT characteristics of patients eventually included in the study.In the modeling group,Mann-Whitney U test and Chi-square test were used to analyze the CT quantitative and qualitative features of ground glass nodule in invasive adenocarcinoma group and non-invasive adenocarcinoma group,and statistically significant indicators in the univariate analysis were selected for multivariate logistic regression analysis to finally determine the independent risk factors of invasive adenocarcinoma.Variance Inflation Factor(VIF)was used to evaluate the degree of collinear interference among independent risk factors.Independent sample t test and Chi-square test were used to analyze the quantitative and qualitative CT features of the GGN of the modeling group and the validation group,and to compare the differences in various imaging features of ground glass nodules between the modeling group and the validation group.3.R software was used to build a nomogram model to predict the invasiveness of GGN,and the validation group was used for external validation of the model;The AUC of the modeling group and the validation group were used to evaluate the model differentiation.Hosmer-Lemeshow goodness of fit test and calibration curve were used to evaluate the fit degree and calibration degree of the model.The decision analysis curve(DCA)of modeling group and validation group was used to evaluate the clinical practicability of the model.Meanwhile,Bootstrap method was used for internal validation.Results:1.Among the 702 patients who met the inclusion and exclusion criteria,female patients accounted for 74.2%,non-smoking patients accounted for 81.1%,and patients < 60 years old accounted for 82.2%;A total of 748 ground glass nodules were included,including 555 in the modeling group and 193 in the validation group,There were 422 ground glass nodules in the upper lobe of both lungs,accounting for 56.4% of the total nodules;Among 748 ground glass nodules,12% were in situ adenocarcinoma,43% were microinvasive adenocarcinoma,and 34.9% were lepidic predominant adenocarcinoma in the invasive adenocarcinoma group.2.Univariate analysis in the modeling group showed that there were significant differences in maximum diameter,mean CT value,consolidation/tumor ratio(CTR),lobulation sign,maximum CT value,pleura traction sign,spiculation sign and vascular convergence sign between the invasive and non-invasive adenocarcinoma groups(P < 0.01).Binary logistic regression analysis showed that maximum diameter,proportion of solid components,maximum CT value,mean CT value,spiculation sign and vascular convergence sign were independent risk factors for the diagnosis of invasive adenocarcinoma with GGN(P < 0.05).Collinearity test results showed that VIF was all less than 5,indicating that there was no collinearity relationship among the above independent risk factors.There was no significant difference in imaging features of ground glass nodules between the modeling group and the validation group(P > 0.05).3.Based on the above independent risk factors,the nomogram prediction model of invasiveness of ground glass nodule was constructed.The AUC values in the modeling group and validation group were 0.910(95%CI: 0.885-0.934)and 0.902(95%CI: 0.859-0.944),respectively,indicating good model differentiation.Hosmer-Lemeshow goodness of fit test of modeling group and verification group indicated that the model had good fitting effect(P > 0.05).The calibration curves of the modeling group and the verification group showed that the risk of invasive adenocarcinoma predicted by the model was in good agreement with the actual risk.The decision analysis curves in the modeling group and validation group showed that the model had good clinical practicability.The AUC value of the Bootstrap internal validation method was 0.905,indicating that the model still had high differentiation ability in the internal validation.Conclusion(s):1.This study shows that Ground-glass nodule lung adenocarcinoma mainly occurs in non-smoking females < 60 years old,the upper lobe of both lungs is common;microinvasive adenocarcinoma is the main pathological subtype in the non-invasive adenocarcinoma group,and lepidic predominant adenocarcinoma in the invasive adenocarcinoma group is the main pathological subtype.2.Maximum diameter,CTR,mean CT value,maximum CT value,spiculation sign and vascular convergence sign were independent predictive factors for the occurrence of IAC in ground glass nodules,which could be used to distinguish invasive and non-invasive adenocarcinoma.3.Combined with quantitative and qualitative features of CT imaging,a nomogram prediction model can be built to predict the invasiveness of ground glass nodules.This model has good prediction efficacy for the invasiveness of GGN,and can provide help for the clinical management and decision-making of GGN.