| 韩雷,单奔,柳勇,于冬洋,周寒松.弥散加权成像直方图参数模型预测高级别胶质瘤复发时间[J].中国医学影像技术,2021,37(6):847~851 |
| 弥散加权成像直方图参数模型预测高级别胶质瘤复发时间 |
| Diffusion weighted imaging histogram parameters model for predicting recurrence time of high-grade glioma |
| 投稿时间：2020-07-08 修订日期：2021-05-15 |
| DOI：10.13929/j.issn.1003-3289.2021.06.011 |
| 中文关键词: 胶质瘤 弥散磁共振成像 直方图 预后 |
| 英文关键词:glioma diffusion magnetic resonance imaging histogram prognosis |
| 基金项目: |
|
| 摘要点击次数: 2808 |
| 全文下载次数: 622 |
| 中文摘要: |
| 目的 评估弥散加权成像（DWI）直方图参数预测模型对预测高级别胶质瘤（HGG）复发时间（TTR）的价值。方法 收集39例经手术病理确诊并术后复发HGG患者，根据TTR分为短期组（TTR≤6个月，n=17）及长期组（TTR>6个月，n=22）；回顾分析术前头部DWI，提取病灶的直方图参数，包括均值（mean）、方差（variance）、偏度（skewness）、峰度（kurtosis）和第1、10、50、90、99百分位数（pere.1%、pere.10%、pere.50%、pere.90%、pere.99%），观察病灶囊变直径是否>30 mm、瘤周水肿、边界清晰与否及有无花环样强化，比较上述参数组间差异。分别基于直方图参数和联合病灶形态特征建立预测HGG的TTR逻辑回归（LR）预测模型及联合预测模型，以受试者工作特征（ROC）曲线评估其诊断效能。结果 长期组囊变直径大于短期组（P<0.05）；variance、skewness、pere.50%值均低于短期组（P均<0.05）。variance预测HGG TTR的效能最高，其诊断敏感度、特异度、准确率及曲线下面积（AUC）分别为70.60%、72.70%、71.80%及0.78。LR预测模型预测HGG TTR的敏感度、特异度、准确率及AUC分别为76.40%、68.20%、71.80%、0.80，联合预测模型分别为76.40%、77.30%、76.90%、0.82。结论 基于DWI直方图参数的LR预测模型对预测HGG的TTR具有一定价值；联合病灶形态特征有助于提高预测准确率。 |
| 英文摘要: |
| Objective To investigate the value of prediction model based on diffusion weighted imaging (DWI) histogram parameters in forecasting the time to recurrence (TTR) of high-grade gliomas (HGG). Methods Totally 39 HGG patients confirmed by postoperative pathology and recurrent after operation were collected and divided into short-term group (TTR≤6 months, n=17) and long-term group (TTR>6 months, n=22) according to TTR. Preoperative head DWI were retrospectively analyzed, and histogram parameters of lesions on DWI were extracted, including the mean, variance, skewness, kurtosis, and the 1st, 10th, 50th, 90th, 99th percentile (pere.1%, pere.10%, pere.50%, pere.90%, pere.99%). The morphologistic characteristics of lesions were recorded for whether the diameter of cystic degeneration >30 mm, with or without peritumoral edema, clear boundary and garland enhancement, and then were compared between groups. Logistic regression (LR) prediction model and combined prediction model for predicting TTR of HGG were established based on histogram parameters or morphological characteristics of lesions combined with histogram parameters, respectively. Receiver operating characteristic (ROC) curve was used to evaluate the diagnostic efficacy of the models. Results Compared with short-term group, long-term group had larger diameter of cystic degeneration (P<0.05) and lower variance, skewness and pere.50% values (all P<0.05). Variance had the best performances for predicting TTR of HGG. The sensitivity, specificity, accuracy and area under the curve (AUC) of variance for predicting TTR of HGG was 70.60%, 72.70%, 71.80% and 0.78, respectively, of LR prediction model in predicting TTR of HGG was 76.40%, 68.20%, 71.80% and 0.80, of the combined prediction model was 76.40%, 77.30%, 76.90% and 0.82, respectively. Conclusion LR prediction model based on DWI histogram parameters had certain value in predicting TTR of HGG. Combining with the morphological characteristics of lesions could help to improve the prediction accuracy of TTR. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
