Objective To evaluate the correlations between CT features and histopathologic subtypes of lung adenocarcinomas presenting as pure ground-glass nodules (pGGN) of 1 cm or less in maximal diameter. Methods CT appearances, pathology and clinical data of 95 patients (97 lesions) who underwent curative resection of lung adenocarcinomas presenting as pGGN≤1 cm in diameter from March 2011 to February 2015 were retrospectively analyzed. Of the 97 lung adenocarcinomas, there were 19 atypical adenomatous hyperplasia (AAH) (19.6%), 31 adenocarcinoma in situ (AIS) (31.9%), 19 minimally invasive adenocarcinoma (MIA) (19.6%) and 28 invasive pulmonary adenocarcinoma (IPA) (28.9%). Fifty (51.5%) were preinvasive (AAH+AIS) and 47 (48.5%) were invasive (MIA+IPA). Lesion size and density were compared among pathologic subtypes using analysis of variance (ANOVA). Lesion size were compared between preinvasive and invasive lesions using 2?independent samples t?test. Lesion location, presence of bubble?like sign, air bronchogram, vessel changes, margin, and tumor?lung interface were compared among histopathologic subtypes using chi?square test. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the cut-off point of size in discriminating preinvasive lesions from invasive lesions. Results Of the 97 lesions, there were no statistically significant differences among histopathologic subtypes in terms of lesion density, presence of bubble?like sign, air?bronchogram, and margin (P>0.05). Mean size of AAH, AIS, MIA and IPA was (0.72 ± 0.19), (0.82 ± 0.14), (0.84 ± 0.11) and (0.85 ± 0.16) cm respectively. There were statistically significant differences among histopathologic subtypes in terms of lesion size (F=3.16, P=0.028). The vessel changes occurred in 2 of AAH, 11 of AIS, 10 of MIA and 17 of IPA. There were statistically significant differences among histopathologic subtypes in terms of vessel changes (χ2=13.22, P=0.004). Lesions with clear tumor?lung interface were in 10 of AAH, 24 of AIS, 17 of MIA, and 26 of IPA. There were statistically significant differences among histopathologic subtypes in terms of tumor?lung interface (χ2=12.67, P=0.005). The optimal cutoff value of lesion size for differentiating preinvasive lesions from invasive lesions was 0.82 cm (sensitivity, 61.7%;specificity, 62.0%). Conclusion Lesion size, vessel changes, and lung?tumor interface may indicate the invasiveness of lung adenocarcinoma presenting as pGGNs of≤1 cm in diameter.