Value of CT Characteristics in Predicting Invasiveness of Adenocarcinoma Presented as Pulmonary Ground-Glass Nodules.

THORACIC AND CARDIOVASCULAR SURGEON(2017)

引用 25|浏览9
暂无评分
摘要
BackgroundLess invasive adenocarcinomas (LIAs) of the lung, including adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA), are indications of sublobar resection and has a 5-year disease-free survival rate of almost 100% after surgery. By distinguishing invasive adenocarcinoma from LIA with computed tomography (CT) characteristics, it is possible to determine the extent of resection and prognosis for patients with ground-glass nodules (GGNs) before surgery. MethodsWe reviewed CT and pathological findings of 728 GGNs in 645 consecutive patients who received curative lung resection in a single center. Only AIS, MIA, and invasive adenocarcinoma were included. Characteristics of CT, including maximum diameter of the lesion (L-max) and maximum diameter of the consolidation (C-max), were assessed thoroughly. ResultsMultivariate logistic regression showed that larger L-max (p<0.001) and nonsmooth margin (p=0.001) were independent factors for invasive adenocarcinoma in pure GGNs (pGGNs). The optimal cut-off value of L-max was 12.0 mm. In mixed GGNs (mGGNs), multivariate analysis revealed that larger L-max (p<0.001), larger C-max (p=0.032), and vacuole sign (p=0.007) were predictive factors for invasive adenocarcinoma, and the area under curve of regression model was 0.866. The optimal cut-off values of L-max and C-max were 15.4 and 5.8mm, respectively. No node metastasis was found in 295 patients who had at least three stations of mediastinal lymph nodes dissected. ConclusionIn pGGNs, larger L-max (>12.0 mm) and nonsmooth margin were reliable predictors for invasive adenocarcinoma. In mGGNs, lesions with larger L-max (>15.4 mm), larger C-max (>5.8 mm), and vacuole sign were more likely to be invasive adenocarcinoma.
更多
查看译文
关键词
computed tomography,ground-glass nodules,lung adenocarcinoma
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要