| Part 1 Interobserver Agreement and Accuracy in Classifications and Measurements of Pulmonary Subsolid Nodules with Different Window Settings and Dimensions?Objective? To compare the interobserver agreement and accuracy in classifications and measurements of SSNs with different window settings and dimensions.And to explore the correlation between quantitative features and pathologic grade?Materials and Methods? We retrospectively evaluated preoperative chest HRCT and pathological data of 157 patients with 159 surgically resected lung adenocarcinoma manifesting as SSNs.According to the pathology,all specimens were divided into two groups: solid-component group and non-solid-component group.Observer 1 and observer 2 independently assessed the presence of solid component and classified all SSNs with lung window setting,mediastinal window setting and-300 threshold semiautomatic segmentation,Cohen's Kappa was calculated to determine interobserver agreements for nodule classifications.Inconsistent opinions of observer 1 and observer 2 decided by observer 3,Mc Memar ?2 and receiver operating characteristic curve were calculated to determine accuracy for nodule classifications.Observer 1 used computer aided diagnosis software to measure the 1D-WNLW,2D-WNLW,1D-SCLW,2D-SCLW,1D-SCMW,2D-SCMW,3D-WNLW,3D-SCLW,3D-SCMW and 3D-SCT of all SSNs.Observer 2 randomly selected 50 SSNs and repeated all the measurements.Intraclass correlation coefficients was calculated to determine interobserver agreements for nodule measurements.The correlations between quantitative features and pathologic grade were evaluated by Spearman's rank correlation.A P value of less than 0.05 indicated statistical significance.?Results? A total of 54 SSNs in solid-component group,including 32 AAH and 22 AIS;a total of 105 SSNs in non-solid-component group,including 47 MIA and 58 IAC.The interobserver agreement in classifications and measurements of SSNs with different window settings and dimensions were substantial or excellent(0.71<k<0.9,0.71<ICC),among which the interobserver agreement and accuracy with-300 threshold semiautomatic segmentation was the best(Kappa=0.831,ICC=0.983,Sens=85%,Spec=61%,PPV=81%,NPV=67%,AUC=0.750).There was moderate positive correlation between quantitative features and pathologic grade(0.4?r<0.7).?Conclusion? The interobserver agreement and accuracy in classifications and measurements of SSNs with-300 threshold semiautomatic segmentation is the best of all.There was positive correlation between all the quantitative features and pathologic grade.Part 2Section 1 Predictive Value of Whole Nodule Size and Solid Component Size of Pulmonary Subsolid Nodule in Three Different Dimensions for the Pathologic Grade?Objective? To compare the predictive value of whole nodule size and solid component size of lung adenocarcinoma manifesting as SSN in three different dimensions with different window settings for pathologic grade.?Materials and Methods? We retrospectively evaluated preoperative chest HRCT and pathological data of 125 patients with 127 SSNs surgically resected and pathologically confrmed lung adenocarcinomas.According to the 5-year disease-free survival,all specimens were divided into two groups: group A included AIS and MIA;group B only included IAC.A doctor with 5 years of experience in imaging diagnosis used computer aided diagnosis software to measure the 1D-SCLW?2D-SCLW?1D-SCMW?2D-SCMW?1D-WNLW?2D-WNLW and 3D-SCT of all SSNs.Mann-Whitney U test was used to compare the difference between all the quantitative features and pathologic grade.Receiver operating characteristic analyses were conducted for the quantitative features.All the quantitative features were evaluated by using univariate logistic regression analysis,significant quantitative features identified by univariate logistic regression analysis were included in the multivariate logistic regression.A P value of less than 0.05 indicated statistical significance.?Results? All specimens were divided into two groups: a total of 69 SSNs in group A,including 22 AIS and 47 MIA;a total of 58 SSNs in group B,only including IAC.1D-SCLW?2D-SCLW?1D-SCMW?2D-SCMW ?1D-WNLW?2D-WNLW and 3D-SCT of group B were significantly larger than those of group A(P<0.0001).ROC analyses indicated that the diagnostic efficiency of 3D-SCT for the pathologic grade was the highest among 7 CT features(AUC=0.887,sensitivity: 81%,specificity: 93%);The cut-off values of 1D-SCLW?2D-SCLW?1D-SCMW?2D-SCMW?1D-WNLW?2D-WNLW and 3D-SCT were 17.50mm?14.75 mm?9.50 mm?7.75 mm?0.50 mm?1.25 mm and 139.00 mm3.Multiple logistic regression analysis revealed that 3D-SCT was the independent predictor of pathologic grade(OR=4.978,95%CI=1.430~17.331,P=0.012).3D-SCT of 139.00 mm3 or greater was a significant indicator of AIC(AUC=0.887,sensitivity:81%,specificity:93%).A P value of less than 0.05 indicated statistical significance.?Conclusion? Among whole nodule size and solid component size of lung adenocarcinoma manifesting as pulmonary subsolid nodule in three different dimensions with different window settings,3D-SCT is found to be the independent predictor of pathologic grade,which is more reliable than one-dimensional and two-dimensional maximum diameter of SSN and solid component.Part 2Section 2 Predictive Value of Solid Component Size and Proportion of Pulmonary Subsolid Nodule with Different Window Settings for the Pathologic Grade?Objective? To compare the predictive value of whole nodule volume,solid component volume and solid component proportion of pulmonary SSNs with different window settings for pathologic grade.?Materials and Methods? We retrospectively evaluated preoperative chest HRCT and pathological data of 125 patients with 127 surgically resected lung adenocarcinoma manifesting as SSNs.According to the 5-year disease-free survival,all specimens were divided into two groups: group A included AIS and MIA;group B included IAC.A doctor with 5 years of experience in imaging diagnosis used computer aided diagnosis software to measure the 3D-WNLW,3D-SCLW,3D-SCMW and 3D-SCT of all SSNs and calculated P-SCLW,P-SCMW and P-SCT.The interobserver agreement regarding quantitative features were evaluated by using intraclass correlation coefficient.Mann-Whitney U test was used to compare the difference between all the quantitative features and pathologic grade.All the quantitative features were evaluated by using univariate logistic regression analysis,significant quantitative features identified by univariate logistic regression analysis were included in the multivariate logistic regression.Receiver operating characteristic analyses were conducted for the quantitative features that exhibited significant differences in the multivariate logistic regression.A P value of less than 0.05 indicated statistical significance.?Results? All specimens were divided into two groups: a total of 69 SSNs in group A,including 22 AIS and 47 MIA;a total of 58 SSNs in group B,only including IAC.The 3D-WNLW?3D-SCLW?3D-SCMW?3D-SCT?P-SCLW?P-SCMW and P-SCT of group B were significantly larger than those of group A(P<0.0001).The univariate logistic regression analysis indicated that 3D-WNLW?3D-SCLW?3D-SCMW?3D-SCT?P-SCLW?P-SCMW and P-SCT were significant(P<0.0001),the multivariate logistic regression analysis indicated that only P-SCT was the independent predictive factor(OR=1.093,95%CI:1.047~1.141,P<0.0001).P-SCT of 6.00% or greater was a significant indicator of AIC(AUC=0.846,sensitivity:79%,specificity:75%).A P value of less than 0.05 indicated statistical significance.?Conclusion? Among whole nodule volume,solid component volume and solid component proportion of pulmonary SSNs with different window settings,P-SCT is found to be the independent predictor of pathologic grade,which can provide more reliable reference for the choice of surgical methods compared with whole nodule volume and solid component volume. |
PDF Full Text Request