Ultra-high Resolution Computed Tomography Accurate Diagnosis Of Pulmonary Pure Ground-glass Nodule

Part ? Ultra-high Resolution Computed Tomography Accurate Diagnosis of Pulmonary Pure Ground-glass NodulePurpose: To retrospectively identify the ability of physiological ventilation-assisted ultra-high resolution Computed Tomography(UHRCT)targeted scan technology(G-protocol)to help differentiate persistent pure ground-glass opacity nodule(PGGN).Materials and Methods: This retrospective study was approved by the institutional review board with waiver of patients' informed consent.324 pathologically confirmed PGGNs,in 270 patients who had undergone G-protocol scan between January 2015 and December 2016,were reviewed including 17 cases of focal interstitial fibrosis(FIF),59 of atypical adenomatous hyperplasia(AAH),154 of adenocarcinomas in situ(AIS),53 of minimally invasive adenocarcinoma(MIA)and 40 of invasive adenocarcinoma(IA).CT characteristics of each nodule were analyzed including the maximum diameter,representative attenuation,homogeneity,shape,margin,tumor-lung interface,air bronchogram,bubble lucency,pleural retraction,and vascular convergence sign.All data were measured on the maximal section of the pulmonary nodules in modified multiplanar reformation images.All the cases were performed ultra-high-resolution computed tomography before operation.To seek the independent predictor in differentiating invasive lesions from preinvasive lesions,logistic regression analysis was conducted.The diagnosis value of logistic regression model was evaluated by using receiver operating characteristic(ROC)curve analysis.For the differentiation of preinvasive lesions from invasive lesions,the optimal cut-off values of the maximum diameter and representative attenuation were calculated by using ROC curve analysis.For the differentiation of AAH from AIS,MIA from AIS,IA from MIA,the optimal cut-off values of the maximum diameter and representative attenuation were calculated by ROC curve analysis.Results: All parameters were significantly differences among the four histopathologic subtypes(P<0.05).Upon further comparison between any 2 groups in size,there were significant differences between AAH and AIS,AAH and MIA,AAH and IA,AIS and MIA and AIS and IA(P<0.008).The optimal cut-off size for AAH from AIS was less than 8 mm(sensitivity,61.0%;specificity,79.9%).The optimal cut-off size for MIA from AIS was larger than 11 mm(sensitivity,50.9%;specificity,78.6%).The optimal cut-off size for IA from MIA was larger than 16 mm(sensitivity,35.0%;specificity,79.9%).All parameters,except shape and tumor-lung interface,all showed significant differences between preinvasive lesions and invasive lesions(P<0.05).The maximum diameter and representative attenuation of invasive lesions was larger than preinvasive lesions(P<0.05).Irregular margin,air bronchogram,bubble lucency,pleural retraction,and vascular convergence were common seen in invasive lesions.Multivariate logistic analysis revealed that the maximum diameter,representative attenuation and irregular margin were the significant differentiator of invasive lesions from preinvasive lesions(P<0.05),with moderate differentiating accuracy(area under ROC curve,0.801).The optimal cut-off of logistic model for invasive lesions was larger than 0.346(sensitivity,66.7%;specificity,80.8%).ROC curve analysis showed the optimal cut-off of one-dimensional maximum diameter for invasive lesions was larger than 10 mm(sensitivity,66.7%;specificity,72.8%).The optimal cut-off of representative attenuation for invasive lesions was larger than-498HU(sensitivity,57.0%;specificity,68.1%).Conclusion: The maximum diameter,representative attenuation and irregular margin were the significant differentiator of invasive lesions from preinvasive lesions.The optimal cut-off of the maximum diameter and representative attenuation for invasive lesions was larger than 10 mm and-498 HU,respectively.Compared to representative attenuation,the maximum diameter gains higher efficiency for evaluating Invasiveness.The image based on the G-protocol can effectively help differentiate PGGN and can be used in clinical application.Part 2 Comparison of Measuring Methods of the Size and Density of Pure Ground-glass NodulePurpose: To retrospectively identify the Application value of different measuring methods of the size and density of pulmonary nodules and its value in the CT accurate diagnosis of lung neoplasm presented with pure ground-glass nodule(PGGN).Materials and Methods: 324 pathologically confirmed PGGNs,in 270 patients who had undergone G-protocol scan between January 2015 and December 2016,were reviewed including 17 cases of focal interstitial fibrosis(FIF),59 of atypical adenomatous hyperplasia(AAH),154 of adenocarcinomas in situ(AIS),53 of minimally invasive adenocarcinoma(MIA)and 40 of invasive adenocarcinoma(IA).The measuring methods of the size included the one-dimensional maximum diameter,two-dimensional and three-dimensional average diameter of the nodule.The measuring methods of density included mean attenuation,representative attenuation,background lung attenuation,and relative attenuation(background lung attenuation minus mean attenuation).All data were measured on the maximal section of the pulmonary nodules in initial sectional images and modified multiplanar reformation images.The changes of the diameter before and after the three-dimensional orthogonality were compared.The difference between the size and density of the preinvasive lesions and invasive lesions were compared.Receiver operating characteristic(ROC)analyses were conducted to determine the optimal measurement method and the optimal cut-off value for differentiating invasive lesions and preinvasive lesions.Results: In initial sectional images,the one-dimensional maximum diameter was larger than two-dimensional and three-dimensional average diameter(P<0.05).There was no significantly differences between two-dimensional and three-dimensional average diameter(P > 0.05).In modified multiplanar reformation images,the one-dimensional maximum diameter was larger than two-dimensional average diameter(P<0.05),the two-dimensional average diameter was larger than three-dimensional average diameter(P<0.05).The one-dimensional maximum diameter,two-dimensional and three-dimensional average diameter of the nodule in modified multiplanar reformation images were larger than that of in initial sectional images(P<0.05).For each measuring methods of size,invasive lesions were larger than preinvasive lesions(P<0.05).Area under ROC curve for the one-dimensional maximum diameter,two-dimensional and three-dimensional average diameter of PGGN in initial sectional images for differentiating invasive lesions from preinvasive lesions was 0.762,0.765 and 0.763,respectively;the optimal cut-off values were all 9mm.Area under ROC curve for the one-dimensional maximum diameter,two-dimensional and three-dimensional average diameter of PGGN in modified multiplanar reformation images for differentiating invasive lesions from preinvasive lesions was 0.764,0.761 and 0.763,respectively;the optimal cut-off values were all 10 mm.The mean attenuation and representative attenuation of invasive lesions was larger than preinvasive lesions(P<0.05).The relative attenuation of preinvasive lesions was larger than invasive lesions(P<0.05).Background lung attenuation showed no significantly differences between invasive lesions and preinvasive lesions(P>0.05).Area under ROC curve for the mean attenuation,representative attenuation and relative attenuation of PGGN for differentiating invasive lesions from preinvasive lesions was 0.764,0.761 and 0.763,respectively;the optimal cut-off values were-591 HU,-498 HU and-299 HU,respectively?Conclusion: The diameter of PGGN in modified multiplanar reformation images were larger than that of in initial sectional images.the one-dimensional maximum diameter,two-dimensional and three-dimensional average diameter of the nodule in initial sectional images and modified multiplanar reformation images can be used for evaluating the invasiveness of PGGN.The one-dimensional maximum diameter of PGGN in modified multiplanar reformation images is recommended.Compared to mean attenuation,representative attenuation and relative attenuation gain higher efficienc for evaluating the invasiveness of PGGN.Representative attenuation is recommended.Part 3 Computed Tomography Diagnosis of Lung Ground-glass Nodules Presented with Reversed Halo SignPurpose: To analyse the computed tomography characteristic and diagnosis of lung ground-glass nodule(GGN)presented with reversed halo sign(RHS).Methods: Thirty-four ground-glass nodules presented with RHS in thirty-four patients(29 females,5 males,age ranged from 24 to 78 yrs,mean 44.2±13.6yrs and median age 42.5yrs)by operation pathology results were included in this study.All data were measured on the maximal section of the pulmonary nodules in modified multiplanar reformation images.CT features of lung nodules including the maximal dimension,the thickness,uniformity and integrity of peripheral ring and density.Density included representive attenuation(the high attenuation of ring,the diameter of region of interest [ROI] larger than 2mm)and intraring attenuation(the attenuation of central region).All the cases were performed ultra-high-resolution computed tomography before operation.Results: There were 32 tumorous lesions,accounting for 94.1%,including 21 preinvasive lesions(7 atypical adenomatous hyperplasia [AAH] and 14 adenocarcinomas in situ [AIS])and 11 invasive lesions(9 minimally invasive adenocarcinoma [MIA] and 2 invasive adenocarcinoma [IA]).2 were benign lesions,accounting for 5.9%.All the lesions were round and well-demarcated.Invasive lesions were significantly larger than preinvasive lesions(9.9±1.6vs8.5±1.8)(P<0.05).The thickness of peripheral ring of invasive lesions were significantly larger than preinvasive lesions(3.9±1.1vs3.1±0.9)(P<0.05).The thickness,uniformity and integrity of peripheral ring,representive attenuation and intraring attenuation showed no significant difference between preinvasive lesions and invasive lesions.The optimal cut-off size for invasive lesions was greater than 9 mm(area under cure [AUC]:0.725;sensitivity,81.8%;specificity,66.7%).The optimal cut-off maximal thickness for invasive lesions was greater than 4 mm(AUC:0.701;sensitivity,45.5%;specificity,90.5%).Conclusion: GGNs presented with RHS strongly suggest the diagnosis of tumor.the lesion size of greater than 9 mm and the maximal thickness of outer rim greater than 4mm can be very specific discriminators of invasive lesions from preinvasive lesions.