基于肿瘤及肿瘤-脑界面纹理特征的胶质母细胞瘤与孤立性脑转移的鉴别

Texture Feature Differentiation of Glioblastoma and Solitary Brain Metastases Based on Tumor and Tumor-brain Interface

DOI 原文链接

用sci-hub下载

ACADEMIC RADIOLOGY

影响因子:3.9

分区:医学2区 / 核医学2区

发表日期:2025 Jan

作者: Yini Chen, Hongsen Lin, Jiayi Sun, Renwang Pu, Yujing Zhou, Bo Sun

DOI: 10.1016/j.acra.2024.08.025

摘要

纹理特征，既来自整个肿瘤区域，也来自肿瘤-脑界面区域，都是区分肿瘤类型及其恶性程度的重要指标。然而，来自这两个区域的纹理特征在鉴别胶质母细胞瘤与转移性肿瘤中的判别价值尚未被充分探索。本研究旨在开发并验证一种结合整个肿瘤区域与10mm肿瘤-脑界面区域纹理特征的诊断模型，以期识别出更稳定、更有效的纹理特征。我们回顾性收集了2010年至2024年间97例胶质母细胞瘤（GBM）和90例单一脑转移（SBM）患者的增强T1加权影像数据。采用机器学习建立多种鉴别GBM与SBM的诊断模型，模型依据整体肿瘤和10mm肿瘤-脑界面区域的纹理特征进行训练。通过5折交叉验证评估模型性能，计算每个模型的受试者工作特征曲线下面积（AUC）。利用Delong检验比较各模型的AUC，并通过Shapley加性解释（SHAP）进一步增强模型的可解释性。所有验证集模型的AUC通过FeAture Explorer（FAE）软件进行比较。在Kruskal-Wallis（KW）和逻辑回归（LR）建立的模型中，采用“一个标准误差”规则获得最高AUC。'10mm_glrlm_GrayLevelNonUniformity'被认为是最稳定且具有预测能力的特征。在训练集、测试集和验证集中，表现最佳的模型并不相同。在测试集中，KW1LR模型的AUC最高，为0.880，准确率为0.824。结合整体肿瘤与肿瘤-脑界面纹理特征的模型有助于区分胶质母细胞瘤与单一转移性肿瘤，而且肿瘤界面纹理特征表现出更高的异质性。

Abstract

Texture features, derived from both the entire tumor area and the region of the tumor-to-brain interface, are crucial indicators for distinguishing tumor types and their degrees of malignancy. However, the discriminative value of texture features from both regions for identifying glioblastomas and metastatic tumors has not been thoroughly explored. The aim of this study is to develop and validate a diagnostic model that combines texture features from the entire tumor area and a 10 mm tumor-to-brain interface region, in an attempt to identify more stable and effective texture features.We retrospectively collected enhanced T1-weighted imaging data from 97 patients with glioblastoma (GBM) and 90 patients with single brain metastasis (SBM) between 2010 and 2024. Machine learning is used to establish multiple diagnostic models for discriminating GBM and SBM based on texture features of the entire tumor and 10 mm tumor-to-brain interface regions. Results underwent evaluation through 5-fold cross-validation analysis, calculating the area under the receiver operating characteristic curve (AUC) for each model. The performance of each model was compared using the Delong test, and the interpretability of the optimized model was further augmented by employing Shapley additive explanations (SHAP).The AUCs for all pipelines in the validation dataset were compared using FeAture Explorer (FAE) software. Among the models established by Kruskal-Wallis(KW) and Logistic Regression(LR), the AUC was highest using the "one-standard error" rule. '10mm_glrlm_GrayLevelNonUniformity' was considered the most stable and predictive feature. The best models in the training set, test set, and validation set were not the same. In the test set, the KW1LR model had the highest AUC of 0.880 and an accuracy of 0.824.The texture feature model that combines the overall tumor and the tumor-brain interface is beneficial for distinguishing glioblastoma from solitary metastatic tumors, and the texture features of the tumor interface exhibit higher heterogeneity.