文献速递神经肿瘤

磁共振成像特征预测脑膜瘤脑侵犯:系统综述与荟萃分析

Magnetic Resonance Imaging Features for Predicting Brain Invasion in Meningiomas: A Systematic Review and Meta-Analysis.
PubMed 文献 · World Neurosurg · 2025-08-05

该系统综述纳入14篇原始研究,评估MRI特征对脑膜瘤脑侵犯(BI)的预测价值;荟萃分析纳入12项研究,汇总AUC为0.91,灵敏度0.85,特异度0.83。亚组分析显示使用≥2个序列、样本量>200例、包含脑-肿瘤界面及ADC可影响诊断效能。

要点

  • MRI对脑膜瘤脑侵犯预测的汇总AUC为0.91,灵敏度和特异度分别为0.85和0.83。
  • 使用≥2个MRI序列较单序列具有更高灵敏度和特异度,加用ADC可进一步提升效能。
  • 样本量>200例及包含脑-肿瘤界面的研究显示更优的灵敏度或诊断比值比。

核心结论

  • MRI对预测脑膜瘤脑侵犯表现出良好的诊断效能。
  • 诊断效能受成像序列、样本量及脑-肿瘤界面特征显著影响。

研究设计与样本

  • 系统综述与荟萃分析,检索Web of Science、PubMed和EMBASE从2016年1月1日至2024年8月1日的原始文章。
  • 系统综述纳入14篇研究脑膜瘤BI的MRI特征的文章;荟萃分析纳入12项原始研究。

主要发现

  • 伴BI的脑膜瘤表现出更高的瘤周水肿体积、不规则肿瘤形状、不完整的脑脊液裂隙征、不均匀对比增强、更大的尺寸、不清晰的脑-肿瘤界面及更低的平均表观扩散系数值。
  • MRI预测BI的汇总AUC为0.91(95%CI 0.88-0.93,SE=0.0165,P=0.0185),汇总灵敏度0.85(95%CI 0.81-0.89,P<0.001),汇总特异度0.83(95%CI 0.76-0.89,P<0.001)。

关键数值与推荐

  • 使用≥2个序列的研究较1个序列的研究灵敏度为0.86 vs. 0.82,特异度为0.85 vs. 0.68;加用ADC后汇总灵敏度和特异度分别为0.89和0.88。
  • 样本量>200例的研究灵敏度为0.86 vs. 0.79,特异度为0.85 vs. 0.76。
  • 包含脑-肿瘤界面的研究灵敏度为0.89 vs. 0.81,诊断比值比为44.3 vs. 20.1,特异度相似(0.84 vs. 0.83)。
  • 使用放射组学的研究特异度为0.85 vs. 0.80,诊断比值比为34.8 vs. 28.9,但灵敏度较低(0.83 vs. 0.91)。

临床怎么用

  • 临床可综合应用多序列MRI(包括ADC)评估脑膜瘤脑侵犯风险。
  • 关注瘤周水肿、肿瘤形状、脑脊液裂隙征、对比增强均匀性、肿瘤大小及脑-肿瘤界面等影像特征。
  • 放射组学分析有助于提高特异度,但需注意其灵敏度可能较低。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVE: To systematically assess magnetic resonance imaging (MRI) features of brain invasion (BI) in meningiomas and to evaluate diagnostic performance of MRI for prediction of BI in meningiomas. METHODS: A comprehensive search was conducted on Web of Science, PubMed, and EMBASE from January 1, 2016, to August 1, 2024, to confirm eligible original articles. Data extracted from the articles included sample size, number of patients with BI or without BI, mean age, male/female ratio, authors, publication year, duration of patient care, study design, strength (tesla) of magnetic field, imaging sequences used, use of radiomics, and reference standard methodology. RESULTS: This systematic review included 14 eligible articles investigating MRI characteristics of BI in meningiomas. Meningiomas with BI exhibited higher volumes of peritumoral edema, irregular tumor shape, incomplete cerebrospinal fluid cleft sign, heterogeneous contrast enhancement, larger sizes, unclear brain-to-tumor interface, and lower mean apparent diffusion coefficient value. The meta-analysis included 12 original studies. The summary area under the curve of MRI for predicting BI in meningiomas was 0.91 (95% confidence interval 0.88-0.93, SE = 0.0165, P = 0.0185), with summary sensitivity and specificity of 0.85 (95% confidence interval 0.81-0.89, P < 0.001) and 0.83 (95% confidence interval 0.76-0.89, P < 0.001), respectively. In subgroup analyses, studies incorporating ≥2 sequences demonstrated superior sensitivity (0.86 vs. 0.82) and specificity (0.85 vs. 0.68)compared with studies incorporating 1 sequence; the addition of apparent diffusion coefficient appeared to further increase diagnostic performance, with summary sensitivity and summary specificity of 0.89 and 0.88, respectively. Studies with a sample size >200 patients had higher sensitivity (0.86 vs. 0.79) and specificity (0.85 vs. 0.76). Studies including brain-to-tumor interface showed better sensitivity (0.89 vs. 0.81) and diagnostic odds ratio (44.3 vs. 20.1), but similar specificity (0.84 vs. 0.83). In addition, studies using radiomics showed better specificity (0.85 vs. 0.80) and diagnostic odds ratio (34.8 vs. 28.9), but exhibited a lower sensitivity (0.83 vs. 0.91). CONCLUSIONS: MRI demonstrated favorable diagnostic efficacy for prediction of BI in meningiomas. Diagnostic performance of MRI was notably influenced by the specific imaging sequences employed, sample size, and characteristics observed at the brain-to-tumor interface.

来源:World Neurosurg,查看原文。版权归原作者所有。

引用

  1. Huang H, Gao Y, Wu L, et al. Magnetic Resonance Imaging Features for Predicting Brain Invasion in Meningiomas: A Systematic Review and Meta-Analysis. World Neurosurg. 2025. PMID: 40754241. doi:10.1016/j.wneu.2025.124336

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。