Article Abstract

Tumor invasiveness defined by IASLC/ATS/ERS classification of ground-glass nodules can be predicted by quantitative CT parameters

Authors: Qian-Jun Zhou, Zhi-Chun Zheng, Yong-Qiao Zhu, Pei-Ji Lu, Jia Huang, Jian-Ding Ye, Jie Zhang, Shun Lu, Qing-Quan Luo


Background: To investigate the potential value of CT parameters to differentiate ground-glass nodules between noninvasive adenocarcinoma and invasive pulmonary adenocarcinoma (IPA) as defined by IASLC/ATS/ERS classification.
Methods: We retrospectively reviewed 211 patients with pathologically proved stage 0-IA lung adenocarcinoma which appeared as subsolid nodules, from January 2012 to January 2013 including 137 pure ground glass nodules (pGGNs) and 74 part-solid nodules (PSNs). Pathological data was classified under the 2011 IASLC/ATS/ERS classification. Both quantitative and qualitative CT parameters were used to determine the tumor invasiveness between noninvasive adenocarcinomas and IPAs.
Results: There were 154 noninvasive adenocarcinomas and 57 IPAs. In pGGNs, CT size and area, one-dimensional mean CT value and bubble lucency were significantly different between noninvasive adenocarcinomas and IPAs on univariate analysis. Multivariate regression and ROC analysis revealed that CT size and one-dimensional mean CT value were predictive of noninvasive adenocarcinomas compared to IPAs. Optimal cutoff value was 13.60 mm (sensitivity, 75.0%; specificity, 99.6%), and −583.60 HU (sensitivity, 68.8%; specificity, 66.9%). In PSNs, there were significant differences in CT size and area, solid component area, solid proportion, one-dimensional mean and maximum CT value, three-dimensional (3D) mean CT value between noninvasive adenocarcinomas and IPAs on univariate analysis. Multivariate and ROC analysis showed that CT size and 3D mean CT value were significantly differentiators. Optimal cutoff value was 19.64 mm (sensitivity, 53.7%; specificity, 93.9%), −571.63 HU (sensitivity, 85.4%; specificity, 75.8%).
Conclusions: For pGGNs, CT size and one-dimensional mean CT value are determinants for tumor invasiveness. For PSNs, tumor invasiveness can be predicted by CT size and 3D mean CT value.