Department of Radiology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China.
Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China.
World Neurosurg. 2023 Jul;175:e264-e270. doi: 10.1016/j.wneu.2023.03.066. Epub 2023 Mar 21.
To investigate the predictive value of noncontrast computed tomography (NCCT) models based on radiomics features and machine learning for early perihematomal edema (PHE) expansion in patients with spontaneous intracerebral hemorrhage (ICH).
We retrospectively reviewed NCCT data from 214 patients with spontaneous ICH. All radiomics features were extracted from volume of interest of hematomas on admission scans. A total of 8 machine learning methods were applied for constructing models in the training and the test set. Receiver operating characteristic analysis and the areas under the curve were used to evaluate the predictive value.
A total of 23 features were finally selected to establish models of early PHE expansion after feature screening. Patients were randomly assigned into training (n = 171) and test (n = 43) sets. The accuracy, sensitivity, and specificity in the test set were 72.1%, 90.0%, and 66.7% for the support vector machine model; 79.1%, 70.0%, and 84.4% for the k-nearest neighbor model; 88.4%, 90.0%, and 87.9% for the logistic regression model; 74.4%, 90.0%, and 69.7% for the extra tree model; 74.4%, 90.0%, and 69.7% for the extreme gradient boosting model; 83.7%, 100%, and 78.8% for the multilayer perceptron (MLP) model; 72.1%, 100%, and 65.6% for the light gradient boosting machine model; and 60.5%, 90.0%, and 53.1% for the random forest model, respectively.
The MLP model seemed to be the best model for prediction of PHE expansion in patients with ICH. NCCT models based on radiomics features and machine learning could predict early PHE expansion and improve the discrimination of identify spontaneous intracerebral hemorrhage patients at risk of early PHE expansion.
研究基于放射组学特征和机器学习的非对比计算机断层扫描（NCCT）模型对自发性脑出血（ICH）患者早期血肿周围水肿（PHE）扩张的预测价值。
我们回顾性分析了 214 例自发性 ICH 的 NCCT 数据。所有放射组学特征均从入院扫描的血肿感兴趣区提取。在训练集和测试集中共应用了 8 种机器学习方法来构建模型。采用受试者工作特征分析和曲线下面积来评估预测价值。
经过特征筛选，共选择 23 个特征来建立早期 PHE 扩张模型。患者被随机分配到训练集（n=171）和测试集（n=43）。在测试集中，支持向量机模型的准确率、敏感度和特异度分别为 72.1%、90.0%和 66.7%；k-近邻模型为 79.1%、70.0%和 84.4%；逻辑回归模型为 88.4%、90.0%和 87.9%；极端梯度提升模型为 74.4%、90.0%和 69.7%；多层感知机（MLP）模型为 74.4%、90.0%和 69.7%；轻梯度提升机模型为 83.7%、100%和 78.8%；随机森林模型为 72.1%、100%和 65.6%。
MLP 模型似乎是预测 ICH 患者 PHE 扩张的最佳模型。基于放射组学特征和机器学习的 NCCT 模型可以预测早期 PHE 扩张，并提高识别有早期 PHE 扩张风险的自发性脑出血患者的鉴别能力。
