【打印本页】 【下载PDF全文】 【HTML】 查看/发表评论下载PDF阅读器关闭

←前一篇|后一篇→

过刊浏览    高级检索

本文已被:浏览 3265次   下载 2879 本文二维码信息
码上扫一扫!
肺部局灶性磨玻璃密度结节多排螺旋CT征象良恶性的Logistic 回归分析
范丽,李清楚,于红,肖湘生,刘士远*
0
(第二军医大学长征医院影像科,上海 200003)
摘要:
目的 分析肺部局灶性磨玻璃密度结节(fGGO)多排螺旋CT(MDCT)影像的良恶性。方法 回顾82例经病理或抗炎治疗后证实的fGGO患者的一般资料(性别、年龄)、病灶部位、大小及MDCT征象(病灶形态、边缘形态、界面、内部结构、邻近结构)。采用Fisher确切概率法、χ2检验、非参数Mann-Whitney U检验及二元Logistic回归分析进行良恶性分析。结果 82例fGGO中恶性61例,良性21例;良、恶性fGGO间一般资料、病灶部位和大小差异均无统计学意义。Logistic回归分析得出:分叶、界面和胸膜凹陷是良恶性fGGO间最重要的鉴别征象,诊断恶性的优势比分别是8.122、3.139、9.076;利用所得回归方程诊断的灵敏度、特异度和准确率分别是93.4%、66.7%、86.6%。结论 肺部局灶性磨玻璃密度结节MDCT影像若表现为具有分叶、清楚毛糙的界面和胸膜凹陷征提示恶性可能性较大。
关键词:  肺结节  螺旋计算机体层摄影术  磨玻璃密度
DOI:10.3724/SP.J.1008.2010.01060
投稿时间:2010-05-12修订日期:2010-09-17
基金项目:国家自然科学基金青年基金(81000602),上海市自然科学基金(10ZR1438900),上海市“登山计划”重大项目(06DZ19503),上海市人才发展基金(026).
Multi-detector CT features of benign and malignant pulmonary focal ground-glass opacity nodules: a binary logistic regression analysis
FAN Li, LI Qing-chu, YU Hong, XIAO Xiang-sheng,LIU Shi-yuan*
(Department of Radiology, Changzheng Hospital, Second Military Medical University, Shanghai 200003, China)
Abstract:
Objective To evaluate the different multi-detector computed tomography features of benign and malignant pulmonary focal ground-glass opacity (fGGO) nodules. Methods The demographic data(sex and age), nodule location, size, and multi-detector CT (MDCT) features (shape, margin, interface, internal characteristics, and adjacent structures) of 82 fGGO patients were retrospectively analyzed. Differences between benign and malignant fGGO were analyzed by using chi-square test, Fisher exact test, and Mann-Whitney U test. Binary logistic regression analysis was also performed to draw a regression equation. Results The 82 fGGO patients included 61 with malignant and 21 with benign nodules. No statistical differences were found between benign and malignant fGGO nodules in terms of demographic data, nodule location, or nodule size. Binary logistic regression analysis showed that lobulation, interface, and pleural indentation were the important features of malignant nodules, with the corresponding odds ratio (OR) being 8.122, 3.139, and 9.076, respectively. The sensitivity, specificity, and accuracy of the obtained equation were 93.4%, 66.7%, and 86.6%, respectively. Conclusion An fGGO nodule with lobulation, well-defined but coarse interface, and pleural indentation in its MDCT image, is more likely to be malignant.
Key words:  pulmonary nodule  spiral computed tomography  ground-glass opacity