Analysis of histopathologic subtypes and CT characteristics of lung adenocarcinomas presenting as pure ground-glass nodule of 1 cm or less in maximal diameter

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. (authors)