| 马惠钰,张巍,侯超,程令刚,张文恺,杨粒芝,何文.术中超声影像组学预测高级别脑胶质瘤异柠檬酸脱氢酶1(IDH1)突变[J].中国医学影像技术,2025,41(4):569~572 |
| 术中超声影像组学预测高级别脑胶质瘤异柠檬酸脱氢酶1(IDH1)突变 |
| Intraoperative ultrasound radiomics for predicting isocitrate dehydrogenase 1 (IDH1) mutation of high-grade glioma |
| 投稿时间：2024-10-12 修订日期：2025-02-27 |
| DOI：10.13929/j.issn.1003-3289.2025.04.013 |
| 中文关键词: 神经胶质瘤 超声检查 异柠檬酸脱氢酶 影像组学 |
| 英文关键词:glioma ultrasonography isocitrate dehydrogenase radiomics |
| 基金项目:国家自然科学基金项目(82271995)。 |
|
| 摘要点击次数: 481 |
| 全文下载次数: 89 |
| 中文摘要: |
| 目的 探讨术中超声影像组学预测高级别脑胶质瘤异柠檬酸脱氢酶1(IDH1)突变的价值。方法 回顾性分析95例高级别胶质瘤(WHO Ⅲ、Ⅳ级),均接受开颅胶质瘤切除术,术中以超声辅助定位肿瘤,后经病理证实诊断。按7[DK(]∶[DK)]3比例划分训练集(n=66,24例IDH1突变型及42例IDH1野生型)与验证集(n=29,11例IDH1突变型及18例IDH1野生型)。于术中超声图像中提取并筛选肿瘤最佳影像组学特征,以随机森林算法建立预测高级别胶质瘤IDH1突变的影像组学模型;绘制受试者工作特征(ROC)曲线,计算曲线下面积(AUC),评估模型预测效能,并采用决策曲线分析(DCA)评估其临床价值。结果 基于术中超声图像共提取851个影像组学特征,最终筛选出5个最佳影像组学特征,以之构建的影像组学模型在训练集和验证集预测高级别胶质瘤IDH1突变的AUC分别为0.902及0.707,差异无统计学意义(P=0.097)。DCA图显示影像组学模型的临床净获益较高。结论 术中超声影像组学能有效预测高级别脑胶质瘤IDH1突变。 |
| 英文摘要: |
| Objective To investigate the value of intraoperative ultrasound radiomics for predicting isocitrate dehydrogenase 1 (IDH1) mutation of high-grade glioma. Methods Ninety-five patients with high-grade glioma (WHO grade Ⅲ and Ⅳ) who underwent craniotomy glioma resection and ultrasound assisted tumor localization during operation and then confirmed by pathology were retrospectively enrolled. The patients were divided into training set (n=66, including 24 IDH1 mutation type and 42 IDH1 wild type) and validation set (n=29, including 11 IDH1 mutation type and 18 IDH1 wild type) at the ratio of 7 ∶ 3. Based on intraoperative ultrasound, radiomics features were extracted, the best ones were screened, and a radiomics model was established for predicting IDH1 mutation of high-grade glioma using random forest algorithm. Receiver operating characteristic (ROC) curve was plotted, the area under the curve (AUC) was calculated to evaluate the predictive efficacy of the model, and decision curve analysis (DCA) was used to evaluate the clinical value of the model. Results A total of 851 radiomics features were extracted based on intraoperative ultrasound, and finally 5 best ones were screened out to construct a radiomics model. The AUC of the radiomics model for predicting IDH1 mutation of high-grade glioma in training and validation sets was 0.902 and 0.707, respectively, with no significant difference (P=0.097). DCA maps showed that the clinical net benefit of the radiomics model was high. Conclusion Intraoperative ultrasound radiomics could effectively predict IDH1 mutation of high-grade glioma. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
