| 崔达华,刘爱连,郭妍,田士峰,牛雅欣,李昕,宋清伟,庄丽娜.基于表观弥散系数图影像组学模型评价子宫内膜样腺癌病理分级[J].中国医学影像技术,2021,37(10):1509~1513 |
| 基于表观弥散系数图影像组学模型评价子宫内膜样腺癌病理分级 |
| Radiomics model based on apparent diffusion coefficient images for differential diagnosis of pathological grade of endometrioid adenocarcinoma |
| 投稿时间：2020-08-27 修订日期：2021-08-09 |
| DOI：10.13929/j.issn.1003-3289.2021.10.018 |
| 中文关键词: 子宫内膜肿瘤 磁共振成像 表观弥散系数 影像组学 |
| 英文关键词:endometrial neoplasms magnetic resonance imaging apparent diffusion coefficient radiomics |
| 基金项目: |
|
| 摘要点击次数: 1272 |
| 全文下载次数: 578 |
| 中文摘要: |
| 目的 观察基于表观弥散系数（ADC）图影像组学模型评价子宫内膜样腺癌（EA）病理分级的价值。方法 纳入97例经术后病理诊断为EA的女性患者，包括中、高分化EA 65例，低分化EA 32例。采用合成少数类过采样技术（SMOTE）使低分化EA样本量达65，之后按7 ∶ 3将患者分为训练集（n=90，中、高分化45例，低分化45例）和验证集（n=40，中、高分化20例，低分化20例）。重建ADC图，提取其影像组学特征；建立Logistic回归模型，绘制受试者工作特征（ROC）曲线、校准曲线和临床决策曲线，评价模型的诊断效能。结果 影像组学模型鉴别训练集和验证集EA病理分级的准确率、曲线下面积（AUC）、敏感度、特异度、阳性预测值（PPV）及阴性预测值（NPV）分别为77.78%、0.87、73.33%、82.22%、80.49%及75.51%和77.50%、0.84、80.00%、75.00%、76.19%及78.95%，诊断效能差异无统计学意义（P=0.66）。结论 基于ADC图的影像组学模型对评价EA病理分级具有一定价值。 |
| 英文摘要: |
| Objective To observe the value of radiomics model based on apparent diffusion coefficient (ADC) images for evaluating pathological grade of endometrioid adenocarcinoma (EA). Methods Ninety-seven female patients with EA diagnosed by postoperative pathology were enrolled, including 65 cases of well-moderate differentiated grade and 32 cases of poor differentiated grade. Synthetic minority oversampling technique (SMOTE) was used to obtain 65 cases of poor differentiated grade EA. Then the cases were divided into training set (n=90, 45 well-moderate differentiated and 45 poor differentiated EA) and verification set (n=40, 20 well-moderate differentiated and 20 poor differentiated EA) at the ratio of 7 ∶ 3. ADC images were reconstructed, and the radiomics were extracted to construct the Logistic regression model. The receiver operating characteristic (ROC) curve, calibration curve and decision curve were drawn to observe the diagnostic efficiency of this model. Results The accuracy, area under the curve (AUC), sensitivity, specificity, positive prediction value (PPV) and negative prediction value (NPV) of the radiomics model for differential diagnosing the pathologic grade of EA in training set and verification set was 77.78%, 0.87, 73.33%, 82.22%, 80.49%, 75.51% and 77.50%, 0.84, 80.00%, 75.00%, 76.19%, 78.95%, respectively. No significant difference of diagnostic efficiencies of this model was found between the two sets (P=0.66). Conclusion Radiomics model based on ADC images had certain value for evaluating pathologic grade of EA. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
