用Logistic回归分析法对直径<2cm孤立肺癌结节CT征象的回顾性分析  

Retrospective analysis of CT manifestations of solitary lung cancer nodules less than 2 cm using Logistic regression analysis

作　　者：林伟添[1] 侍丽[1] 梁知遇 黄健威[1] LIN Weitian;SHI Li;LIANG Zhiyu;HUANG Jianwei(Department of Radiology,the 3rd Affiliated Hospital of Guangzhou Medical University,Guangzhou 510150,China)

机构地区：[1]广州医科大学附属第三医院放射科,广东广州510150

出　　处：《实用放射学杂志》2019年第5期726-729,共4页Journal of Practical Radiology

摘　　要：目的运用Logistic回归分析法回顾性分析直径<2cm肺部孤立结节的CT征象对早期肺癌的诊断意义。方法选取156例病理证实的直径<2cm的肺部小结节病例进行回顾性分析,对各CT征象进行统计学赋值,并进行二分类Logistic回归分析筛选出可能是肺癌结节的CT征象并分析其为早期肺癌的危险度。结果6个CT征象进入Logistic方程模型,依次为“磨玻璃征”(8.12)、“分叶征(6.72)”、“血管纠集征”(6.02)、“毛刺征”(5.07)、“空洞坏死”(3.41)、空泡征(1.02)。结论“磨玻璃征”是肺癌结节危险度最高的征象,以CT征象建立的Logistic模型对鉴别孤立肺癌小结节有较大的参考意义。Objective To analyze the characteristic manifestations and early diagnostic value of CT in solitary pulmonary nodules(SPNs)less than 2 cm using Logistic regression analysis.Methods 156 patients with SPN less than 2 cm confirmed by pathology were collected.Statistical assignment was performed and binary Logistic regression was implemented for CT manifestations.Those features,which might be signs of lung cancer,were extracted from the CT images and their risk levels were also analyzed.Results Six CT signs including"ground glass sign"(8.12),"lobulation sign"(6.72),"vascular convergence sign"(6.02),"spicule sign"(5.07),"necrosis and cavitation"(3.41),and"vacuole sign(1.02)"were enrolled in the Logistic equation.Conclusion"Ground-glass sign"is associated with the highest risk level for lung cancer nodules.The Logistic model constructed from CT manifestations is helpful for identifying solitary lung cancer nodules.

关 键 词：立肺癌结节 LOGISTIC回归分析 计算机体层成像 磨玻璃征 

分 类 号：R563[医药卫生—呼吸系统] R814.42[医药卫生—内科学]