文献速递神经肿瘤

偶然发现脑膜瘤MRI影像组学特征预测肿瘤生长的探索性定量分析

Novel predictors of tumor growth by exploratory quantitative analysis of radiomics features from MRI data for incidentally discovered meningioma.
PubMed 文献 · J Neurooncol · 2025-09-16

本研究针对偶然发现的脑膜瘤(IDM),比较24例生长组与25例未生长组初始MRI T2影像的影像组学特征,基于3D Slicer 5.6.2与PyRadiomics 3.1.0自建半自动批处理t检验程序,在肿瘤、肿瘤外缘、全脑及对侧白质ROI中分析七大分类共107个特征,发现五个分类中10个特征组间差异显著(p<0.05),提示影像组学特征可作为IDM生长的潜在预测因子。

要点

  • 初始MRI T2影像组学可定量评估偶然发现脑膜瘤的生长风险
  • 24例生长组与25例未生长组比较,10个特征差异显著(p<0.05)
  • 结果属探索性,需更大队列与验证后才能用于临床决策

核心结论

  • 影像组学特征可从初始MRI识别偶然发现脑膜瘤生长的潜在预测因子,但尚需更大队列与验证确认临床效用。

研究设计与样本

  • 回顾性比较影像组学分析:24例肿瘤生长病例与25例未生长病例的初始MRI T2影像特征对比。
  • 自建基于3D Slicer 5.6.2与PyRadiomics 3.1.0的半自动批处理程序,对每个特征行t检验。
  • ROI置于肿瘤、肿瘤外缘、全脑及肿瘤对侧白质,共分析七大分类107个特征。

主要发现

  • 五个分类中10个特征在生长组与未生长组间差异显著(p<0.05)。
  • 显著特征包括2个First Order统计特征、2个Shape特征、4个灰度共生矩阵特征、1个灰度大小区域矩阵特征和1个邻域灰度差矩阵特征。

临床怎么用

  • 对偶然发现脑膜瘤患者,可尝试用初始MRI T2影像组学特征辅助评估未来生长风险,但目前仅为研究性工具。

局限性

  • 样本量小(生长组24例、未生长组25例),结果为探索性,需更大队列与外部验证确认。

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

PURPOSE: Predicting future tumor growth from initial imaging of incidentally discovered meningioma (IDM) could inform treatment decisions. However, most factors identified in prior studies on meningioma growth are qualitative. The aim of this study is to identify factors associated with tumor growth using quantitative radiomics features from MRI data. METHODS: MRI T2 features from initial imaging of 24 tumor growth cases were compared with those of 25 cases without growth. An in-house program was developed to reduce the time required for data analysis. This program is based on the open-source software 3D Slicer 5.6.2 and PyRadiomics 3.1.0. It enables semi-automatic batch t-test analyses for each feature to compare tumor growth and non-growth groups. Regions of interest (ROIs) were placed in the tumor, outer tumor edge, whole brain, and white matter contralateral to the tumor. A total of 107 features were analyzed across seven classifications: First Order, Shape, Gray Level Co-occurrence Matrix, Gray Level Run Length Matrix, Gray Level Size Zone Matrix, Gray Level Dependence Matrix, and Neighboring Gray Tone Difference Matrix. A t-test was used to identify significant predictors. RESULTS: Ten features across five classifications showed significant differences (p < 0.05): 2 First Order statistics, 2 Shape features, 4 Gy Level Co-occurrence Matrices, 1 Gy Level Size Zone Matrix, and 1 Neighboring Gray Tone Difference Matrix. CONCLUSIONS: Potential predictors of IDM growth were identified using radiomics features. Future studies with larger cohorts and validation will be essential to confirm the clinical utility and improve the predictive accuracy of these features.

来源:J Neurooncol,查看原文。版权归原作者所有。

引用

  1. Oi Y, Minamoto H, Taniyama I, et al. Novel predictors of tumor growth by exploratory quantitative analysis of radiomics features from MRI data for incidentally discovered meningioma. J Neurooncol. 2025. PMID: 40958042. doi:10.1007/s11060-025-05186-8

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。