Objective:To analyze the correlation between high resolution CT (HRCT ) features of pulmonary pure ground‐glass nodules (PGGN ) including early invasive pulmonary adenocarcinomas and preinvasive lesions and the new pathologic classification of lung adenocarcinoma ,and to evaluate the predictive values of HRCT in the pathologic classification of lung adenocarcinoma with PGGN .Methods:The data of 123 patients hospitalized from January 2014 to June 2014 in a single central and diagnosed by HRCT as early peripheral lung adenocarcinoma or atypical adenoma with PGGN were retrospectively analyzed .The correlation between HRCT morphological characteristics ,size ,and density of PGGN and the 2015‐edition new classification of lung adenocarcinoma were analyzed ,the best predictors were screened out ,and a model was constructed and verified .Results:All of nine image morphological features were significantly correlated with the new pathological classification (Pearson correlation test , P< 0 .05 ) . Among them , eight morphological features including lobulation , spiculation , air/bronchial inflation ,pleural indentation ,edge regularity ,shape regularity ,density uniformity and vessel convergence had a positive linear correlation with the new pathological classification (P<0 .01) .Futhermore ,four continuous variables describing the size and density of the lesion including maximum cross‐sectional area ,lesion length in cranial‐caudal direction ,average actual density of PGGN and the relative average density were significantly correlated with the new pathological classification (P<0 .05) .Multinomial logistic regression analysis was adopted to screen out the seven best predictors :spiculation ,lesion lgenth in cranial‐caudal direction ,average actual density ,clear demarcation of tumor ,gender ,age ,and vessel convergence . After the model was constructed ,and the likelihood ratio test showed that the overall matching rate was 70 .7% ,while the matching rate of atypical adenomatous hyperplasia (AAH ) was up to 92 .9% .Conclusions:The HRCT characteristics of PGGN were significantly correlated with the new 2015‐edition WHO pathologic classification of lung adenocarcinoma .The new pathological classification of PGGN can be predicted by HRCT ,and the best predictors were spiculation ,lesion length in cranial‐caudal direction ,average actual density ,clear demarcation of tumor ,gender ,age ,and vessel convergence .%目的：分析纯磨玻璃密度肺部早期浸润性腺癌和浸润前病变的高分辨率CT （high resolution CT ，HRCT ）特征与肺腺癌病理新分类的相关性，评估H RC T对纯磨玻璃密度肺腺癌病理新分类的预测价值。方法：回顾性分析2014年1月—2014年6月单中心收治的123例H RC T表现为肺部纯磨玻璃密度早期周围型肺腺癌或不典型腺瘤的患者资料。分析纯磨玻璃结节（pure ground‐glass nodules ，PGGN）的HRCT形态学特征、大小和密度等因素与2015版肺腺癌病理新分类的相关性，筛选出最佳预测因素，构建模型并进行验证。结果：9项影像形态学特征均与病理新分类分组明显相关（P＜0．05）。其中，分叶、毛刺、空气／支气管充气、胸膜凹陷、边缘规整、形状规则、密度均匀及血管集束等8项形态学特征与病理新分类分组线性正相关（P＜0．01）；而且最大截面面积、头脚方向长度、PGGN实际平均密度和相对平均密度等4个描述病灶大小与密度的连续计数变量与病理新分类分组显著相关（P＜0．05）。采取多因素回归分析，筛选出了7项最佳预测因素：毛刺征、头脚方向长度、实际平均密度、肺瘤分界清晰性、性别、年龄及血管集束征。构建模型后，似然比检验显示总体匹配率达70．7％，其中对非典型腺瘤样增生（atypical adenomatous hyperplasia ，AAH）的匹配率高达92．9％。结论：PGGN的 HRCT影像学表现与2015版WHO肺腺癌新分类明显相关，HRCT检查可预测PGGN的病理新分类，其中最佳预测因子为毛刺征、头脚方向长度、实际平均密度、肺瘤分界清晰性、性别、年龄及血管集束征。
展开▼
