Objectives: Pulmonary ground-glass nodules (GGNs) are occasionally diagnosed as invasive adenocarcinomas. This study aimed to evaluate the clinicopathological features of patients with pulmonary GGNs to identify factors predictive of pathological invasion.
Methods: We retrospectively evaluated 101 pulmonary GGNs resected between July 2006 and November 2013 and pathologically classified them as adenocarcinoma in situ (AIS; n = 47), minimally invasive adenocarcinoma (MIA; n = 30), or invasive adenocarcinoma (I-ADC; n = 24). The age, sex, smoking history, tumor size, and computed tomography (CT) attenuation of the 3 groups were compared. Receiver operating characteristic (ROC) curve analyses were performed to identify factors that could predict the presence of pathologically invasive adenocarcinomas.
Results: Tumor size was significantly larger in the MIA and I-ADC groups than in the AIS group. CT attenuation was significantly greater in the I-ADC group than in the AIS and MIA groups. In ROC curve analyses, the sensitivity and specificity of tumor size (cutoff, 11 mm) were 95.8% and 46.8%, respectively, and those for CT attenuation (cutoff, -680 HU) were 95.8% and 35.1%, respectively; the areas under the curve (AUC) were 0.75 and 0.77, respectively. A combination of tumor size and CT attenuation (cutoffs of 11 mm and -680 HU for tumor size and CT attenuation, respectively) yielded in a sensitivity and specificity of 91.7% and 71.4%, respectively, with an AUC of 0.82.
Conclusions: Tumor size and CT attenuation were predictive factors of pathological invasiveness for pulmonary GGNs. Use of a combination of tumor size and CT attenuation facilitated more accurate prediction of invasive adenocarcinoma than the use of these factors independently.