扩散峰度成像在乳腺良恶性病变鉴别诊断中的应用：Meta分析  

Application of the diffusion kurtosis imaging for the study of breast lesions：Meta-analysis

作　　者：管雪琴 李红林[1] 方进[1] 蔡湘怡 周全[1] GUAN Xueqin;LI Honglin;FANG Jin;CAI Xiangyi;ZHOU Quan(Medical Imaging Center,the First Affiliated Hospital of Jinan University,Guangzhou 510630,China)

机构地区：[1]暨南大学附属第一医院医学影像中心,广东广州510630

出　　处：《实用放射学杂志》2018年第10期1534-1537,1552,共5页Journal of Practical Radiology

摘　　要：目的 采用Meta分析评价扩散峰度成像（DKI）在乳腺良恶性病变鉴别诊断中的价值。方法 检索Pubmed、Cochrane协作网、Cochrane图书馆、OVID、中国生物医学文献数据库（CBM）、中文科技期刊全文数据库（CNKI）、万方数据知识服务平台、维普中文科技期刊数据库（VIP）自建库以来至2017年3月国内外公开发表的关于DKI在乳腺良恶性病变鉴别诊断中的价值的中英文文献。按照Cochrane协作网推荐的诊断试验纳入标准筛选文献，并对纳入研究进行质量评价后提取相关数据。采用Review Manager 5.3进行数据分析，合并诊断效应量标准化均数差值（SMD）、诊断比值比（DOR）及绘制汇总受试者工作特征曲线（SROC曲线），计算曲线下面积（AUC）。结果 共纳入7篇文献，病灶644个，乳腺恶性病变的平均扩散峰度（MK）明显高于良性病变（SMD 1.66,95%CI 1.21~2.10;Z=7.31,P〈0.000 01），其中5篇文献可提取敏感度、特异度、AUC及截点数据，汇总敏感度（SEN）、特异度（SPE）及DOR分为0.92（95%CI 0.88~0.94）、0.87（95%CI 0.81~0.91）、79.07（95%CI 43.35~144.22），AUC为0.956 2。结论 DKI定量参数MK在乳腺良恶性病变鉴别诊断中有较可靠的应用价值。Objective To evaluate the applying value of diffusion kurtosis imaging（DKI） in the differential diagnoses of breast lesions with Meta-analysis.Methods Published papers in Chinese and English on DKI in the breast lesions were searched from Pubmed,Cochrane collaboration, Cochrane library, OVID, CBM database, China National Knowledge Infrastructure （ CNKI）, Wanfang Chinese periodical database and VIP database from inception to March 2017.And select according to the criteria for diagnostic recommend by Cochrane method group.Quality evaluation was performed for the eligible papers before data extraction.Statistical analysis was performed with software Review Manager 5.3.Standardized mean differenees（SMD）and 95% confidence intervals（CI）were used to compare MK values between malignant lesions and the benign lesions.The diagnosis effect size diagnostic odds ratio（DOR）was calculated and summary receiver operating characteristic（SROC）curve was drawn to obtain the area under curve（AUC）.Results 7 studies with 644 lesions were selected for the Meta-analysis. We found that MK values measured by DKI were significantly higher in malignant lesions than in benign lesions in the breast （SMD 1.66, 95%CI 1.21-2.10;Z= 7.31,P〈0.000 01）. And 5 studies were extracted with sensitivity, specificity,AUC and cut point.The pooled sensitivity, specificity, diagnostic odds ratio were calculated： 0.92 （ 95 %CI 0.88- 0.94 ）, 0.87 （ 95 %CI 0.81 - 0.91 ）, 79.07 （95%CI 43.35 144.22） ,the AUC was 0.956 2.Conclusion MK is a reliable parameter for differentiating malignant lesions from benign lesions in breast.

关 键 词：扩散峰度成像 乳腺病变 平均扩散峰度 META分析 

分 类 号：R445.2[医药卫生—影像医学与核医学] R655.8[医药卫生—诊断学]