| 潘越,党浩丹,孟晓琳,张聪,林宇,王瑞民,徐白萱.11C-蛋氨酸PET/CT影像组学模型评估胶质母细胞瘤异柠檬酸脱氢酶1状态[J].中国医学影像技术,2024,40(6):832~837 |
| 11C-蛋氨酸PET/CT影像组学模型评估胶质母细胞瘤异柠檬酸脱氢酶1状态 |
| 11C-methionine PET/CT radiomics model for evaluating isocitrate dehydrogenase1 status of glioblastoma |
| 投稿时间：2024-02-05 修订日期：2024-03-16 |
| DOI：10.13929/j.issn.1003-3289.2024.06.008 |
| 中文关键词: 胶质母细胞瘤 异柠檬酸脱氢酶 正电子发射断层显像和计算机体层摄影术 影像组学 |
| 英文关键词:glioblastoma isocitrate dehydrogenase positron-emission tomography and computed tomography radiomics |
| 基金项目:国家自然科学基金(82001859)。 |
| 作者 | 单位 | E-mail | | 潘越 | 中国人民解放军医学院, 北京 100853
中国人民解放军总医院第一医学中心核医学科, 北京 100853 | | | 党浩丹 | 中国人民解放军总医院第一医学中心核医学科, 北京 100853 | | | 孟晓琳 | 中国人民解放军医学院, 北京 100853
中国人民解放军总医院第一医学中心核医学科, 北京 100853 | | | 张聪 | 中国人民解放军医学院, 北京 100853
中国人民解放军总医院第一医学中心核医学科, 北京 100853 | | | 林宇 | 中国人民解放军医学院, 北京 100853 | | | 王瑞民 | 中国人民解放军总医院第一医学中心核医学科, 北京 100853 | | | 徐白萱 | 中国人民解放军总医院第一医学中心核医学科, 北京 100853 | xbx301@163.com |
|
| 摘要点击次数: 351 |
| 全文下载次数: 261 |
| 中文摘要: |
| 目的 观察11C-蛋氨酸(MET)PET/CT影像组学模型评估胶质母细胞瘤异柠檬酸脱氢酶1(IDH1)状态的价值。方法 回顾性分析157例接受11C-MET PET/CT检查的胶质母细胞瘤患者资料,包括68例IDH1突变及89例野生型;按8 ∶ 2比例将其分为训练集(n=125)与验证集(n=32)。基于PET/CT图像勾画病灶ROI并提取、筛选影像组学特征,分别建立逻辑回归(LR)、支持向量机(SVM)及决策树(DT)影像组学模型;同时基于患者年龄及影像组学特征绘制列线图;对比观察影像组学模型及临床-影像组学列线图评估IDH1状态的效能。结果 DT影像组学模型评估训练集胶质母细胞瘤IDH1状态的曲线下面积(AUC)为0.910,大于LR(0.697)及SVM(0.698)模型(P均＜0.05)。验证集中,DT模型评估胶质母细胞瘤IDH1状态的AUC为0.805,大于LR模型(0.740)及临床-影像组学列线图(0.704)(P均＜0.05)。结论 基于DT的11C-MET PET/CT影像组学模型有助于评估胶质母细胞瘤IDH1状态。 |
| 英文摘要: |
| Objective To explore the value of 11C-methionine (MET) PET/CT radiomics model for evaluating isocitrate dehydrogenase1 (IDH1) status of glioblastoma. Methods Data of 157 patients with glioblastoma who underwent 11C-MET PET/CT examination, including 68 cases of IDH1 mutation and 89 cases of IDH1 wild-type were retrospectively analyzed. The patients were divided into training set (n=125) and validation set (n=32) at the ratio of 8:2. Based on PET/CT images, lesions ROI were delineated and radiomics features were extracted and screened to establish radiomics models with logistic regression (LR), support vector machine (SVM) and decision trees (DT), respectively. Meanwhile, the nomogram based on patients' age and radiomics features was drawn. The efficacy of radiomics models and clinical-radiomics nomogram for evaluating IDH1 status were comparatively observed. Results The area under the curve (AUC) of DT radiomics model for evaluating IDH1 status of glioblastoma in training set was 0.910, higher than that of LR (0.697) and SVM (0.698) models (both P＜0.05). In validation set, the AUC of DT model for evaluating IDH1 status of glioblastoma was 0.805, which was higher than that of LR model (0.740) and clinical-radiomics nomogram (0.704) (both P＜0.05). Conclusion 11C-MET PET/CT radiomics model based on DT was helpful for evaluating IDH1 status of glioblastoma. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
