MRI弥散加权成像对预测乳腺癌新辅助化学治疗效果的Meta分析

Meta-analysis of neoadjuvant chemotherapy effects for predicting breast cancer by MRI diffusion-weighted imaging

目的系统评价MRI弥散加权成像(DWI)在预测乳腺癌新辅助化学治疗结果方面的准确性。方法选择PubMed、EMBASE和Cochrane library数据库,采用EMTREE术语(用于EMBASE)、医学标题术语(用于Medline)和文本词汇(用于其他),即("breast cancer"或"breast neoplasm")和("diffusion weighted magnetic resonance imaging"或"diffusionweighted MRI"或"diffusion MRI"或"DWI")和("neoadjuvant therapy"或"neoadjuvant treatment")和("chemotherapy"或"pharmacotherapy"或"drug therapy"),进行全面的文献检索。时间为1996年至2018年11月21日。评估纳入研究的方法学质量,对其灵敏度和特异度进行异质性检验并进行合并加权分析,绘制受试者工作特征(ROC)曲线,计算曲线下面积(AUC)。结果最终纳入12篇文献,536例研究对象。Meta分析结果显示:合并加权灵敏度和合并加权特异度及其95%置信区间(CI)分别为85%[95%CI(0.76～0.92)]和78%[95%CI(0.64～0.88)]。异质性分析灵敏度(I^2=53.51,P=0.01)和特异度(I^2=83.18,P <0.001)差异有统计学意义,主要是由于存在阈值效应(r=0.694;P=0.012)。使用ROC曲线,AUC为0.89(95%CI:0.86～0.92)。Deeks漏斗图不对称检验表明,无明显的发表偏倚(P=0.54)。结论 DWI是预测乳腺癌新辅助化学治疗疗效的有价值的成像工具。

Objective To systematically detect the accuracy of diffusion-weighted magnetic resonance imaging(DWI) for predicting treatment outcomes of neoadjuvant chemotherapy in patients with breast cancer. Methods The PubMed, EMBASE and Cochrane library databases were enrolled, and EMTREE terminology(for EMBASE), medical title terminology(for Medline),and text vocabulary(for others), the('breast cancer' or 'breast neoplasm') and('diffusion weighted magnetic resonance imaging' or 'diffusion weighted MRI' or 'diffusion MRI' or 'DWI') and('neoadjuvant therapy' or 'neoadjuvant treatment') and('chemotherapy' or 'pharmacotherapy' or 'drug therapy') were used to comprehensive literature search. The time of studies was from 1996 to November 21, 2018. The methodological quality of included studies was assessed, and the sensitivity, specificity were tested for heterogeneity and combined, the receiver operating characteristic(ROC) curve was plotted to calculate area under curve(AUC). Results Twelve literature with 536 subjects were finally enrolled. Meta-analysis indicated that the combined sensitivity and pool specificity and their 95 % confidence interval(CI) were 85 %(95 % CI 0.76-0.92) and 78 %(95 % CI 0.64-0.88), respectively. Due to the threshold effect(r = 0.694, P = 0.012), there were significant staistically differences between heterogeneities sensitivity(I^2= 53.51, P = 0.01) and specificity(I^2= 83.18, P = 0.01). Using the ROC curve, the AUC was 0.89(95 %CI 0.86-0.92). The Deeks funnel plot asymmetry test showed no obvious publication bias( P = 0.54). Conclusion It is demonstrated that DWI is a valuable imaging tool for predicting the efficacy of neoadjuvant chemotherapy in breast cancer.