GGN样早期肺腺癌CT形态学特征的Fisher判别  

Fisher discrimination of morphological features of GGN sample early lung adenocarcinoma by spiral CT scanning

作　　者：徐佳佳 王红[1] 赵年[2] 刘四斌 Xu Jiajia;Wang Hong;Zhao Nian;Liu Sibin(The Second Clinical Medical College of Yangtze University,Hubei Jingzhou 434020,China;Department of Medical Imaging,DongfengGeneral Hospital,Hubei Shiyan 442000,China)

机构地区：[1]长江大学第二临床医学院,湖北荆州434020 [2]十堰市东风总医院影像科,湖北十堰442000

出　　处：《现代肿瘤医学》2018年第1期109-113,共5页Journal of Modern Oncology

基　　金：湖北省荆州市科技局立项项目(编号:2015042)

摘　　要：目的:探讨依据CT形态学特征建立的Fisher判别函数模型,对磨玻璃结节(ground-glass nodules,GGN)样早期肺腺癌浸润前及浸润性病变进行鉴别诊断的价值。方法:回顾性分析荆州中心医院与东风总医院在2014年9月至2017年2月收治并经病理证实的磨玻璃结节样早期肺腺癌患者160例,通过对浸润前与浸润性病变的多种CT形态学特征的统计分析,通过卡方检验,以两组具有统计学差异的CT特征建立Fisher判别函数公式,然后运用交叉核实法,计算总体误判率及准确率。结果:160例患者分为两组,浸润前病变组53例,浸润性病变组107例,毛刺征、分叶征、结节形状、瘤-肺界面、平均直径、胸膜凹陷征、血管集束征及实性成分有无具有统计学差异,并以其为判别指标建立Fisher判别公式Z=2.046X1+2.006X2+2.015X3-1.473X4+0.189X5+0.023X6+1.813X7+1.957X8-1.687。误判率为6.3%,准确率是93.7%。结论:Fisher判别模型对鉴别诊断GGN样早期肺腺癌浸润前与浸润性病变具有较高的准确性及临床实用价值。Objective: To investigate the use of Fisher discriminant function model that is established on the basis of CT morphological features,for lesion before infiltration and invasive lesions of the ground-glass nodules( GGN)sample early lung adenocarcinoma. Methods: A retrospective study was conducted in 160 cases of GGN sample early lung adenocarcinoma. These cases in Jingzhou central hospital and Dongfeng general hospital from September 2014 to February 2017 were confirmed by pathology. The CT morphological features were evaluated by chi-square test and Fisher discriminant was applied to parameter with significant difference. Cross-validation was used for estimating the rate of miscarriage justice. Results: 160 patients with GGN were divided into two groups,the lesion before infiltration group( n = 53),invasive lesions group( n = 107). Spicule signs,lobulation sign,nodules form,nodule interface,average diameter,pleural indentation sign,vascular convergence sign and presence of solid components were significant difference. The formula of fisher discriminant was Z = 2. 046 X1 + 2. 006 X2 + 2. 015 X3-1. 473 X4 + 0. 189 X5+ 0. 023 X6 + 1. 813 X7 + 1. 957 X8-1. 687. The rate of miscarriage of justice was 6. 3%,the accuracy was 93. 7%.Conclusion: Fisher discriminant mode for differential diagnosis GGN sample early lung adenocarcinoma before infiltration and invasive lesions had high accuracy and the clinical practical value.

关 键 词：肺磨玻璃结节 早期肺腺癌 浸润前病变 浸润性腺癌 体层摄影术 X线计算机 

分 类 号：R730.44[医药卫生—肿瘤] R734.2[医药卫生—临床医学]