外周血甲基化Septin-9检测对结直肠癌诊断价值的Meta分析

Diagnostic value of circulating methylated Septin-9 as a biomarker for colorectal cancer: a Meta-analysis study

目的:Meta系统性评价Septin-9在结直肠癌中的诊断价值。方法:检索Pub Med、生物医学外文全文服务系统、Ovid、中国学术期刊全文数据库、中国生物医学文献数据库、维普中文科技期刊数据库、万方数据库,日期截止至2017年1月。严格按照纳入和排除标准筛选文献,并依据QUADAS标准对文献进行质量评估。使用双变量Meta模型合并检验效能分析,通过Q检验、I2检验检测研究异质性,通过亚组分析、敏感性分析探讨异质性来源,通过Deek's漏斗图评估发表偏移。结果:共纳入11篇文献;Septin-9诊断结直肠癌的合并敏感度为0.70(95%CI:0.67~0.72),合并特异度为0.91(95%CI:0.90~0.92),诊断比值比为28.76(95%CI:17.70~46.75),SROC的曲线下面积(AUC)为0.922 1。异质性检验提示不存在阈值效应,但存在其他因素引起的异质性。通过敏感性分析发现纳入文献中有一个离群值。亚组分析结果显示,根据1/3判为阳性和2/3判为阳性分组:AUC为0.939 7比0.826 5。根据亚洲人种和欧美人种分组:AUC为0.936 8比0.921 0。Deek's漏斗图评估纳入文献不存在发表偏倚。结论:外周血甲基化Septin-9诊断结直肠癌的整体准确性较高,可作为一项很好的指标用于结直肠癌诊断。

Objective： A meta-analysis was use to systematically assess the diagnostic value of Septin-9 in Colorectal cancer. Methods： Literature fulfilling the criteria was searched in Pub Med,Foreign Medical Journals Platform,Ovid,CNKI,CBM,Wan Fang and VIP Databases from inception to Jan. 2017. Literatures were strictly screened according to the inclusion and exclusion criteria. Study quality was assessed in terms of the Quality Assessment of Diagnostic Accuracy Studies（ QUADAS） checklist. A bivariate Meta-analysis model was employed to assess the pooled accuracy,and study heterogeneity was evaluated via Cochran-Q and I2tests;subgroup analysis and sensitivity analysis were conducted to deeply trace the sources of heterogeneity; publication bias was judged by Deek＇s funnel plot. Results： A total of 11 studies were included. Analysis of methylated Septin-9 achieved a pooled sensitivity of 0. 70（ 95% CI： 0. 67-0. 72）,specificity of 0. 91（ 95% CI： 0. 90-0. 92）,and DOR of 28. 76（ 95% CI： 17. 70-46. 75）,corresponding to an AUC of 0. 922 1. Heterogeneity test suggested that there was obvious heterogeneity from non-threshold effect. Sensitivity analysis identified one outlier study. Subgroup analysis results showed that the AUC of 1/3 positive to 2/3 positive group was 0. 939 7 versus0. 826 5,and the AUC of the Asian population group to the Caucasian population group was 0. 936 8 versus 0. 921 0. Funnel plot（ Deek＇s） revealed no publication bias. Conclusion： Our data indicate that circulating methylated Septin-9 seemed to harbor a relatively high accuracy in conforming colorectal cancer,and might be popularized as a routine biomarker for colorectal cancer detection.