单、双指数模型DWI鉴别肺部良恶性病变:Meta分析

目的 采用Meta分析评价单、双指数模型DWI鉴别诊断肺部良恶性病变的效能。方法 检索Embase、Pubmed、Cochrane Library、CNKI、万方和维普数据库建库至2018年12月，单指数及双指数模型DWI诊断肺部良恶性病变相关文献，筛选符合纳入标准者进行质量评价，采用Meta-Disc 1.4及Stata 15.0软件进行统计分析。结果 共纳入10篇文献（中、英文文献各5篇）、672例患者（695个病灶）。ADC值、D值诊断肺部良恶性病变的敏感度分别约0.78[95% CI（0.73，0.83）]、0.90[95% CI（0.87，0.93）），特异度分别约0.71[95% CI（0.64，0.78）]、0.64[95% CI（0.57，0.70）]，诊断比值比分别为12.59[95% CI（4.93，32.11）]、19.58[95% CI（7.06，54.29）]，汇总受试者工作特征曲线AUC分别为0.848 9、0.885 1。结论 ADC值、D值鉴别肺良恶性病变均有高诊断效能，D值诊断效能优于ADC值。

DWI based on monoexponential and biexponential models in differential diagnosis of benign and malignant lung lesions: Meta-analysis

Objective To evaluate the diagnostic efficacy of DWI based on monoexponential and biexponential models in differential diagnosis of benign and malignant lung lesions. Methods Studies related to diagnosis of benign and malignant lung lesions by DWI based on monoexponential and biexponential models were retrieved from Embase, Pubmed, Cochrane Library, CNKI, Wanfang Med Online and VIP databases which were performed up to December, 2018. The studies that meet the inclusion criteria were evaluated with quality assessment of diagnostic accuracy studies. Meta-Disc version 1.4 and STATA 15.0 software were used for statistics analysis. Results Totally 10 studies (5 English, 5 Chinese), including a total of 695 lesions in 672 patients were enrolled. The sensitivity of ADC value and D value in differential diagnosis of benign and malignant lung lesions was 0.78 (95%CI[0.73, 0.83]), 0.90 (95%CI[0.87, 0.93]), specificity was 0.71 (95%CI[0.64, 0.78]), 0.64 (95%CI[0.57, 0.70]), the diagnostic odds ratio was 12.59 (95%CI[4.93, 32.11]), 19.58 (95%CI[7.06, 54.29]), and the AUC of summary ROC curve was 0.848 9, 0.885 1, respectively. Conclusion ADC value and D value both have good diagnostic efficacy in distinguishing benign and malignant lung lesions, and the diagnostic efficacy of D value is better than ADC value.