MR扩散加权成像诊断脑梗死的Meta分析

目的 评价MR扩散加权成像(DWI)诊断脑梗死的临床价值及纳入研究质量.方法 检索美国医学索引(Medline)、Ovid、Springer、Elsevier、Lippincott Williams & Wilkins(LWW)数据库及中国期刊网国内文献库(CNKI)1997年1月至2007年1月公开发表的中英文文献,按照Cochrane协作网推荐的诊断试验的纳入标准筛选文献;采用Meta分析软件对纳入文献汇总分析,获得汇总敏感度和特异度及其95%可信区间,并绘制汇总受试者工作特征(SROC)曲线、森林图和倒漏斗图.结果 按照纳入标准,共获取研究DWI诊断超急性期(发病时间≤6 h)脑梗死的文献12篇,DWI诊断超急性期脑梗死的汇总敏感度、特异度、诊断比数比及其95%可信区间分别为92%、87%和180.37,SROC曲线下面积(Az)为0.9717.敏感度分析显示去除低年资医师组诊断影响后,高年资组评价研究间异质性差异无统计学意义(Q=4.34,P>0.05),提示DWI诊断超急性期脑梗死的价值受到低年资医师判断组影响.倒漏斗图不对称提示存在发表偏倚,发表偏倚系数(Z)=10.0824.按照纳入标准,共获取研究DWI诊断超急性期和急性期(发病时间≤24 h)脑梗死文献共15篇,去除低年资医师组的诊断影响后,DWI诊断急性期脑梗死的汇总敏感度、特异度、诊断比数比及其95%可信区间分别为92%、91%和623.62,Az为0.9659.敏感度分析显示组间异质性差异无统计学意义(Q=2.70,P>0.05),提示研究人群无明显异质性.倒漏斗图不对称提示存在发表偏倚,但较超急性组减低(Z=12.6194).结论 通过Meta分析和存在的偏倚分析提示,尽管可以认为DWI作为一种无创性检查方法诊断急性和超急性脑梗死具有较高的敏感度、特异度和准确度,但是纳入文献存在偏倚.

MR diffusion weighted imaging detecting cerebral infarction:a Meta-analysis

Objective To determine the diagnostic value of diffusion-weighted imaging(DWI)in hyperacute and acute cerebral infarction by using Meta-analysis.Methods Based on validity criteria for diagnostic research published by the Cochrane Methods Group on Screening and Diagnostic,studies in English and Chinese from 1 997 to 2007 were selected from Medline,Cochrane,Springer,Ovid,Elsevier,LWW and CNKI(China National Knowledge Infrastructure).The characteristics of the included articles were appraised and extracted. Statistical analysis was performed with the software Meta-test 0.6 and Comprehensive meta-analysis 2.0.Heterogeneity of the included articles was tested.which was used to select proper effect model to calculate pooled weighted values of sensitivity and specificity and the corresponding 95% CL Summary receiver operating characteristic(SROC)curve was performed and the area under the curve(Az)was calculated.Publication bias was analyzed by Funnel Plot in Comprehensive Meta.analysis 2.0.A sensitivity analysis was performed.Results Twelve articles meeting inclusion criteria were analyzed for the value of DWI in hyperacute cerebral infarction.The pooled seusitivitv.specificity and diagnostic odds ratio was 92%,87%,180.37 respectively,Az=0.9717.Novice is a main factor for total diagnostic effect(Q=4.34,P>0.05).Non-asymmetric funnel plot suggested the publication bias.Fifteen articles meeting inclusion criteria were analyzed for the value of DWI in both hyperacute and acute cerebral infarction(≤24 h).The pooled sensitivity,specificity diagnostic odds ratio was 92%、91%. 623.62 respectively.Az=0.9659.Fixed effects model used in Meta-analysis for database suggested homogeneity(Q=2.70,P>0.05).Nonasymmetrie funnel plot suggested the publication bias.Conclusions As a noninvasive method,diffusion-weighted imaging is valuable in detecting hyperacute and acute cerebral infarction.More support from multi-center prospective researches is desirable.