Luo Jie, Cao Yong, Nian Weiqi, Zeng Xiaohua, Zhang Hui, Yue Yeli, Yu Feng
Chongqing Cancer Institute.
Med Ultrason. 2018 Feb 4;1(1):43-49. doi: 10.11152/mu-1209.
The purpose of this diagnostic meta-analysis was to explore the benefit of shear wave elastography (SWE) combinedly used with conventional ultrasound (US) in the diagnosis of the benign and malignant breast lesion.
After a literature search on MEDLINE, Cochrane library, and Embase we included 14 studies with a total 1951 patients and 2060 breast lesions for further analyses. Summary descriptive statistics such as pooled sensitivity, specificity, and summary receiver operating characteristics curve were generated via a bivariate random effect model. Summary indicators such as area under curve (AUC) and confident region were used to compare the performance of conventional US and combination of conventional US and shear wave imaging (SWI) or two-dimensional and three-dimensional SWI.
As indicated by the results, the pooled sensitivity, specificity, and diagnostic odds ratio for combined usage of SWE and the conventional US were 0.877 (95%CI: 0.855-0.896) and 0.849 (95%CI: 0.826-0.869) and 40.164 (95%CI:31.135- 51.811). The AUC for combined use and the conventional US only were 0.928 and 0.899, suggesting a promise of integrating SWE in the routine of breast lesion examination. Also, the summary AUC for 2D and 3D SWE were 0.917 and 0.952 respectively. No significant difference was found between 2D and 3D SWE. No obvious publication bias was identified after employing Deeks' regression test for asymmetry.
Our analysis concluded that SWE improved the differential diagnosis of the breast lesion.
本诊断性荟萃分析旨在探讨剪切波弹性成像（SWE）联合传统超声（US）在乳腺良恶性病变诊断中的益处。
在对MEDLINE、Cochrane图书馆和Embase进行文献检索后，我们纳入了14项研究，共计1951例患者和2060个乳腺病变进行进一步分析。通过双变量随机效应模型生成汇总描述性统计数据，如合并敏感度、特异度和汇总受试者工作特征曲线。使用曲线下面积（AUC）和置信区间等汇总指标来比较传统超声以及传统超声与剪切波成像（SWI）联合或二维和三维SWI的性能。
结果表明，SWE与传统超声联合使用的合并敏感度、特异度和诊断比值比分别为0.877（95%CI：0.855 - 0.896）、0.849（95%CI：0.826 - 0.869）和40.164（95%CI：31.135 - 51.811）。联合使用和仅使用传统超声的AUC分别为0.928和0.899，这表明在乳腺病变检查常规流程中整合SWE具有前景。此外，二维和三维SWE的汇总AUC分别为0.917和0.952。二维和三维SWE之间未发现显著差异。采用Deeks回归检验不对称性后，未发现明显的发表偏倚。
我们的分析得出结论，SWE改善了乳腺病变的鉴别诊断。
