| 王友红,柳勇,韩婷婷,韩雷.基于MR对比增强T1WI纹理分析鉴别泪腺淋巴瘤与泪腺炎性假瘤[J].中国医学影像技术,2022,38(6):837~841 |
| 基于MR对比增强T1WI纹理分析鉴别泪腺淋巴瘤与泪腺炎性假瘤 |
| Texture analysis based on MR contrast-enhanced T1WI for distinguishing lacrimal lymphoma and lacrimal inflammatory pseudotumor |
| 投稿时间：2021-11-01 修订日期：2022-03-08 |
| DOI：10.13929/j.issn.1003-3289.2022.06.011 |
| 中文关键词: 泪腺 淋巴瘤 炎性假瘤 磁共振成像 纹理分析 |
| 英文关键词:lacrimal gland lymphoma inflammatory pseudotumor magnetic resonance imaging texture analysis |
| 基金项目: |
|
| 摘要点击次数: 1117 |
| 全文下载次数: 283 |
| 中文摘要: |
| 目的 观察基于MR对比增强T1WI （CE-T1WI）纹理分析鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的价值。方法 回顾性分析经病理证实的21例泪腺淋巴瘤（淋巴瘤组）和25例泪腺炎性假瘤（炎性假瘤组）的眼眶MRI表现，基于CE-T1WI提取病灶直方图、灰度共生矩阵、灰度游程矩阵、绝对梯度、自回归模型和小波变换6种共279个纹理特征参数，采用组间比较、组内相关系数（ICC）及最小绝对收缩和选择算子（LASSO）回归筛选最佳纹理特征，建立核函数分别为线性核（LK）、多项式核（PK）和径向基函数核（RBFK）的支持向量机（SVM）分类模型，筛选最优核函数。针对最佳纹理特征及组间差异有统计学意义的MRI表现，以最优核函数建立联合模型；并以受试者工作特征（ROC）曲线评估各模型鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的效能。结果 相比炎性假瘤组，淋巴瘤组病灶边界更清晰、强化更均匀（P均<0.01），组间其余MRI表现差异无统计学意义（P均>0.05）。共201个纹理特征组间差异有统计学意义，经筛选10个最佳纹理特征用于建立SVM分类模型，其中PK为最优核函数，相应SVM分类模型鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的效能最佳，其敏感度、特异度、准确率及曲线下面积（AUC）分别为90.47%、88.00%、89.13%及0.93，联合模型分别为95.23%、92.00%、93.47%及0.96；联合模型与最优SVM分类模型的AUC差异无统计学意义（P=0.33）。结论 基于MR CE-T1WI纹理分析可有效鉴别泪腺淋巴瘤与泪腺炎性假瘤。 |
| 英文摘要: |
| Objective To investigate the value of texture analysis based on MR contrast-enhanced T1WI (CE-T1WI) for differentiating lacrimal lymphoma and lacrimal inflammatory pseudotumor. Methods Orbital MRI data of 21 patients with lacrimal lymphoma (lymphoma group) and 25 patients with lacrimal inflammatory pseudotumor (inflammatory pseudotumor group) confirmed by pathology were retrospectively analyzed. Totally 279 feature parameters of lesions, including histogram, gray-level co-occurrence matrix, gray-level run length matrix, absolute gradient, autoregressive mode and wavelet transform were extracted based on CE-T1WI. Then the optimum texture features were screened by using comparison between groups, intra-class correlation coefficient (ICC) and least absolute shrinkage and selection operator (LASSO) regression. And support vector machine (SVM) classification models with kernel function of linear kernel (LK), polynomial kernel (PK) and radial basis function kernel (RBFK) were established, respectively, to screen the optimal kernel function, which was used to establish a combined model based on MRI manifestations being significant different between groups and the optimum texture features. Finally, receiver operating characteristic (ROC) curves were drawn to evaluate the efficacies of models for differential diagnosis of lacrimal lymphoma and lacrimal inflammatory pseudotumor. Results Compared those in inflammatory pseudotumor group, lesions in lymphoma group had clearer boundary and more uniform enhancement (both P<0.01), while no significant difference of other MRI features was found between groups (all P>0.05). Totally 201 texture features were statistical different between groups, among which 10 optimum texture features could be used to establish SVM classification models. PK showed the optimal kernel function, and the corresponding SVM classification model had the best performance for differentiating lacrimal lymphomas and lacrimal inflammatory pseudotumor, with the sensitivity, specificity, accuracy of 90.47%, 88.00%, and 89.13%, respectively, and the area under the curve (AUC) of 0.93. The sensitivity, specificity, accuracy of combined model was 95.23%, 92.00% and 93.47%, respectively, and AUC was 0.96. There was no significant difference in AUC between the combined model and the best SVM classification model (P=0.33). Conclusion Texture analysis based on MR CE-T1WI could effectively distinguish lacrimal lymphoma and lacrimal inflammatory pseudotumor. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
