Department of Joint and Sports Medicine, Qujing First People's Hospital, Qujing, Yunan Province, China.
Department of Medical Imaging Center, Qujing First People's Hospital, Qujing, Yunan Province, China.
Medicine (Baltimore). 2022 Nov 4;101(44):e31574. doi: 10.1097/MD.0000000000031574.
This meta-analysis aimed to explore the diagnostic value of diffusion kurtosis imaging (DKI) compared to mono-exponential diffusion weighted imaging (DWI) in the diagnosis of breast cancer.
A systematic electronic literature search (up to September 2020) was conducted for published English-language studies comparing the diagnostic values of DKI and DWI for the detection of breast cancer. The data of mean kurtosis (MK), mean diffusivity (MD), and apparent diffusion coefficient (ADC) were extracted to construct 2 × 2 contingency tables. The pooled sensitivities, specificities, and areas under the receiver operating characteristic curve (AUCs) were compared between DKI and DWI in the diagnosis of breast cancer.
Eight studies were finally included, with a total of 771 patients in the same population. Pooled sensitivities were 82.0% [95% confidence interval (95% CI), 78.2-85.3%] for ADC, 87.3% (95% CI, 83.9-90.1%) for MK, and 83.9% (95% CI, 80.2-87.1%) for MD. Pooled specificities were 81.1% (95% CI, 76.7-84.9%) for ADC, 85.1% (95% CI, 81.1-88.5%) for MK, and 83.2% (95% CI, 79.0-86.8%) for MD. According to the summary receiver operator characteristic curve analyses, the AUCwas 0.901 for ADC, 0.930 for MK, and 0.918 for MD (ADC vs MK, P = .353; ADC vs MD, P = .611). No notable publication bias was found, while significant heterogeneity was observed.
Although DKI is feasible for identifying breast cancer, MD and MK offer similar diagnostic performance to ADC values. Thus, we recommend that DKI should not be included in the routine evaluation of breast lesions now.
本荟萃分析旨在探讨扩散峰度成像（DKI）与单指数扩散加权成像（DWI）相比在诊断乳腺癌中的诊断价值。
对截至 2020 年 9 月发表的比较 DKI 和 DWI 检测乳腺癌诊断价值的英文文献进行系统的电子文献检索。提取平均峰度（MK）、平均弥散度（MD）和表观弥散系数（ADC）的数据，以构建 2×2 四格表。比较 DKI 和 DWI 在诊断乳腺癌中的汇总敏感度、特异度和受试者工作特征曲线（ROC）下面积（AUC）。
最终纳入 8 项研究，共有 771 例患者在同一人群中。ADC 的汇总敏感度为 82.0%（95%置信区间（95%CI），78.2-85.3%），MK 为 87.3%（95%CI，83.9-90.1%），MD 为 83.9%（95%CI，80.2-87.1%）。ADC 的汇总特异度为 81.1%（95%CI，76.7-84.9%），MK 为 85.1%（95%CI，81.1-88.5%），MD 为 83.2%（95%CI，79.0-86.8%）。根据汇总受试者工作特征曲线分析，ADC 的 AUC 为 0.901，MK 为 0.930，MD 为 0.918（ADC 与 MK 比较，P=.353；ADC 与 MD 比较，P=.611）。未发现明显的发表偏倚，但存在显著的异质性。
尽管 DKI 可用于识别乳腺癌，但 MD 和 MK 的诊断性能与 ADC 值相似。因此，我们建议目前不应将 DKI 纳入乳腺病变的常规评估。
