Abstract
BACKGROUND: Subsolid pulmonary nodules are common computed tomography (CT) findings of primary lung adenocarcinoma. It is of clinical value to determine the clinical treatment strategies based on CT features. The aim of this study is to find the valuable CT characteristics on differential diagnosis and the degree of invasion prediction by a retrospectively analysis of three groups subsolid nodules, including benign, and invasive adenocarcinoma. METHODS: The CT findings of 106 cases of resected sub-solid nodules were retrospectively analyzed. The nodules were firstly divided into benign and malignant groups and the malignant group was further divided into non/micro-invasive group (atypical adenomatous hyperplasia/adenocarcinoma in situ/minimally invasive adenocarcinoma) and invasive adenocarcinoma group. The nodule size, proportion of solid components, tumor-lung interface, shape, margin, pleural traction, air bronchus sign, vascular abnormalities inside the nodule were evaluated. The univariate analysis (χ2 test, non-parametric test Mann-Whitney U test) was performed to screen statistically significant variables and then enrolled in further multivariate Logistic regression analysis. RESULTS: Multivariate logistic regression analysis showed that a clear tumor-lung interface, air bronchus sign, and pulmonary vascular abnormalities were important indicators of malignant nodules with hazard ratios of 38.1 (95%CI: 5.0-287.7; P<0.01), 7.9 (95%CI: 1.3-49.3; P=0.03), 7.2 (95%CI: 1.4-37.0; P=0.02), respectively. The proportion of solid components was the only significant indicator for identifying invasive adenocarcinoma from AAH/AIS/MIA , with a risk ratio of 1.04 (95%CI: 1.01-1.06, P=0.01). CONCLUSIONS: SSNs with clear tumor-lung interface, air bronchus sign, and pulmonary vascular abnormality inside nodule are more likely to be malignant. A higher percentage of solid components indicates a higher likelihood to be an invasive lesion in malignant SPNs.