The Value Of Dual Energy Computed Tomography For Distinguishing Pathological Subtype Of Lung Adenocarcinoma Manifesting As Ground Glass Nodule

Objective To evaluate quantitative parameters generated from Dual-energy computed tomography(DECT)combined with CT features in distinguishing pathological subtypes of lung adenocarcinoma manifesting as ground glass nodule(GGN).Methods We retrospectively enrolled 87 resected GGN lung adenocarcinoma with DECT examinations between December 2017 and March 2019 were collected in the Third Affiliated Hospital of Soochow University.The imaging data were analyzed retrospectively.According to the pathological subtypes of lung adenocarcinoma,the included GGN were divided into following two groups:invasive adenocarcinoma group(n=62);pre-invasive adenocarcinoma and minimally invasive adenocarcinoma group(n=25).CT features including the maximal diameter(Dmax),the margin,internal bronchial morphology,internal vascular morphology,whether there was bubble lucency,pleural indentation,vascular convergence,the CT value on enhanced monochromatic of 40 keV～190 keV(CT 40keV-190keV)and the slope of the energy spectrum decay curve between the two groups were compared by univariate analysis.Binary logistic regression analysis was used to establish a predictive model,and the odds ratio(OR)of independent predictors were calculated.The diagnostic performance of different parameters was compared by receiver operating characteristic curves and Z test.Results The proportion of margin,internal bronchial morphology,and internal vascular morphology,present of vascular convergence were higher in IAC group than in PIA-MIA group,as well as the Dmax,and CT 40keV-190keV of GGN(P range:0.001～0.020).Dmax,internal vascular morphology and CT 60keV were independent predictors(OR=30.921,6.750,50.361,P<0.05).The combined prediction model was build,and the area under the curve(AUC)of it was 0.922 with the accuracy of 88.5%.The AUC of the combined predictive model was significantly higher than the AUC of Dmax,internal vascular morphology,and CT 60keV alone between the IAC group and PIA-MIA group(all P<0.05).Conclusion(1)The Dmax,the proportion of margin,internal bronchial morphology,and internal vascular morphology,present of vascular convergence and CT 40keV-190keV of invasive adenocarcinoma group were higher than that of pre-invasive adenocarcinoma and minimally invasive adenocarcinoma group.(2)The Dmax,internal vascular morphology and CT 60keV of GGN are independent predictors of pathological subtypes of lung adenocarcinoma.(3)CT features combined with quantitative parameters of DECT is an effective method to distinguish pathological subtypes of lung adenocarcinoma manifesting as GGN,which has a high diagnostic efficiency.