3D及2D ADC直方图在鉴别肺孤立性实性病变中的价值  

Histogram Analysis of Apparent Diffusion Coefficient Map Differentiate Malignant and Benign Disease in Lung Based on 3D and 2D ROI

作　　者：包盈莹 李新春[1] 万齐[1] 邹乔[1] 邓颖诗[1] 雷永霞[1] 郑晓涛[1] 余煜栋 雷强[1] BAO Yingying;LI Xinchun;WAN Qi(Department of Radiology,The First Affiliated Hospital of Guangzhou Medical University,Guangzhou 510120,P.R.China)

机构地区：[1]广州医科大学附属第一医院放射科

出　　处：《临床放射学杂志》2018年第8期1293-1297,共5页Journal of Clinical Radiology

基　　金：国家自然基金青年项目(编号:81601457);广东省医学科研基金项目(编号:A2016410);广州市科技计划项目(编号:20150010234);广州市医药卫生科技项目(编号:2015A11066)

摘　　要：目的探讨全瘤（3D）及面积（2D）表观扩散系数（ADC）直方图在鉴别肺部孤立性实性病变中的价值。方法前瞻性搜集经CT检查发现肺部孤立性实性病变患者70例,良性病变19例,恶性肿瘤51例,治疗前行常规MRI、DWI（b=600 mm/s^2）检查,利用Omni-Kinetics软件生成2D及3D ADC直方图并记录其参数。采用MannWhitney检验比较各参数,并运用受试者工作特征曲线（ROC）分析其对鉴别肺部良恶性病变的诊断效能。结果恶性组2D及3D的ADC最小值、最大值、中位数、平均值均低于良性组（P〈0.05）,3D及2D的ADC中位数（AUC=0.842、0.809）、平均值（AUC=0.834、0.807）有较好诊断效能。当3D ADC中位数取1.400×10^-3mm^2/s时,诊断的准确率最高,敏感度为74.5%,特异度为89.50%。恶性组的偏度（P〈0.05）及峰度（P〈0.01）均较良性组高。3D及2D的ADC 50th、75th、90th有显著意义（P〈0.001）。诊断效能较好的参数3D ROI的AUC均较2D ROI大。结论 3D ADC直方图可以较好地区分肺部良恶性实性病变,其中ADC中位数诊断效能最高。Objective To explore the role of histogram analysis of apparent diffusion coefficient（ ADC） maps in differentiating malignancy from benign pulmonary diseases and to evaluate the diagnostic performance of ADC maps based on 3D and 2D ROI. Methods Seventy patients with solid pulmonary lesions detected by chest CT were included. There were 51 malignant lesions and 19 benign lesions which were all confirmed by pathology. All patients were performed conventional MRI scans,DWI MRI scans（ b = 0 and 600 s/mm^2）. Regions of interest containing the lesions were drawn on every section and the maximum section of the ADC map to derive ADC histogram parameters. Each parameter was compared by using the Mann-Whitney U test. Receivers operating characteristic（ ROC） curves were constructed to determine the optimum threshold for each histogram parameter,and diagnostic value of 3D and 2D ROI was compared. Results Malignant lesions demonstrated lower minimum,maximum,median,mean values of ADC than benign lesions（ P 0. 050）. ROC curve analysis ofthe median（ AUC = 0. 842,0. 809,respectively） and mean values（ AUC = 0. 838,0. 807,respectively） of 3D and 2D ADC yielded a better diagnostic performance. Median of 3D ADC has the best accuracy rate with a cutoff value of 1. 400 × 10^-3 mm^2/s whose sensitivity is 74. 5%,and specificity is 89. 5%. The kurtosis（ P 0. 050） and skewness（ P 0. 010） were higher in malignant diseases. The 50 th,75 th,90 th of 3D ADC and 90 th of 2D ADC showed significant differences between malignant and benign diseases（ P 0. 001）. The 75 th ADC had the best diagnostic performance which shows significant difference from 5 th and 25 th of ADC. AUC of 3D ADC was larger than 2D ADC. Conclusion Histogram analysis of 3D ADC maps can be a useful tool for discriminating the malignant from benign pulmonary diseases which is better than that of 2D ADC maps,and the medium has the highest diagnostic value.

关 键 词：肺实性病变 表观扩散系数 直方图 

分 类 号：R445.2[医药卫生—影像医学与核医学] R734.2[医药卫生—诊断学]