Differential Diagnostic Value Of Multi-slice Computed Tomography On Analyzing The Pure Ground Glass Nodules Of Minimal Invasive Adenocarcinomas And Non-invasive Neoplasms

Objective: To retrospectively investigate the differentiating multi-slice computed tomographic(MSCT) features between minimally invasive adenocarcinoma(MIA) and pre-invasive adenocarcinoma(adenocarcinoma in situ, AIS, and atypical adenomatous hyperplasia,AAH) appearing as pure ground glass nodules. We aimed to analyze the MSCT findings of pure ground glass nodules diagnosed pathologically as MIA and pre-invasive adenocarcinoma in order to investigate whether the volume and the mean CT attenuations enable distinction of them.Materials and Methods: From June 2012 to December 2014, 78 patients with 81 pure GGNs identified on thin section CT scans, then underwent multiplanar reconstruction(MPR) was performed at First Affiliated Hospital of Dalian Medical University. There were 13 males and 65 females, the age ranged from 17 to 79 years with average age of 56.15±1.32 years. In our study, patients with localized and pure GGN without anysolid component were resected completely. Among the 81 pathologically proved pure GGNs, 20 were AAH, 31 were AIS and 30 were MIA.The CT thin section images( slice thickness/slice interval 1.25mm) were sent to the GE-ADW 4.5 workstation. All pure GGNs were underwent multiplanar reconstruction( coronal and sagittal). Volume rendering software was used to post-process the original images, the larger than 5mm p GGN were calculated by CT histogram to analysis the volume and the mean CT attenuations.The MSCT findings that were analyzed for each lesion included lesion location, size, shape(round, oval, polygonal, irregular), margin(smooth, shallow notch, lobular, speculated), tumor-lung interface(clear or unclear), the relationship between p GGN and blood vessels(I, vessels passing by or through p GGN; II, intact vessels passing through p GGN with tiny branches; III, distorted, dilated or tortuous vessels seen within p GGN; IV, more complicated vasculature other than described above), internal characteristics and surrounding malignant signs(bubble sign, air bronchogram, pleural tag notch). Two observers measure and record the volume and the mean CT attenuations. The result was evaluated by ICC.MSCT appearances of p GGN were analyzed by using Fisher exact test. Lesion size using analysis of variance(AVOVA) and ROC curves. The volume using analysis of Mann-Whitney U test. The volume of pre-invasive and minimally invasive lesions were assessed using ROC curves. The mean CT attenuations using analysis of two independent t test. Statistical significance was evaluated with software( SPSS, version 17.0). A P value less than 0.05 was considered to indicate a statistically significant difference.Results: Among the MSCT findings, there were significant difference in shape(P<0.01), margin(P<0.01), tumor-lung interface(P=0.04,P<0.05), the relationship between p GGN and blood vessels(P<0.01), internal characteristics(P=0.003,P<0.05)and surrounding malignant signs(P=0.016,P<0.05). No statistical differences were found between pre-invasive and minimally invasive lesions in term of location(P=0.332,P>0.05). The inter observer agreement for size scores is very good(ICC value > 0.75). The lesion size between pre-invasive and minimally invasive lesions was significant difference(One-Way ANOVA,F=15.652,P<0.01).The size of pre-invasive lesions was less than that of MIA(P<0.05). ROC analysis showed that the area under the ROC curve(AUC) for lesion size was 0.771 and that the optimal cutoff value of lesion size for differentiating pre-invasive lesions from MIA was more than 7.8mm(sensitivity, 70%; specificity, 72.5%). The inter observer agreement for the volume scores is very good(ICC value > 0.75).There were significant difference between pre-invasive lesions and MIA in the volume. ROC analysis showed that the area under the ROC curve(AUC) for lesion volume was 0.889 and that the optimal cutoff value of lesion volume for differentiating pre-invasive lesions from MIA was more than 0.37 cm3(sensitivity, 71.4%; specificity, 94.4%). The inter observer agreement for the mean CT attenuations scores is very good(ICC value > 0.75).No statistical differences were found between pre-invasive and minimally invasive lesions in the mean CT attenuations(P=0.46,P>0.05).Conclusion: The MSCT appearances can help predict and diagnose the minimally invasive adenocarcinoma and pre-invasive adenocarcinoma. The volume of p GGN can improve the diagnoses and identify the minimally invasive adenocarcinoma and pre-invasive adenocarcinoma.