Radiomics features on non-contrast computed tomography predict early enlargement of spontaneous intracerebral hemorrhage
Radiomics features on non-contrast computed tomography predict early enlargement of spontaneous intracerebral hemorrhage
复制标题
DOI：
10.1016/j.clineuro.2019.105491
复制
发表时间：
2019-10-01
影响因子：
1.9
通讯作者：
Jiang, Xiaoli
中科院分区：
文献类型：
--
作者：
Li, Hui;Xie, Yuanliang;Jiang, Xiaoli
目的：探讨非对比ct （NCCT）放射组学特征对自发性脑出血（siich）早期扩大的预测价值。患者和方法：167例SICH患者根据24小时随访CT图像血肿体积为基线NCCT的30%和/或6ml分为血肿扩大组和非血肿扩大组。将所有病例的基线NCCT图像导入放射组学软件，提取初始血肿的放射组学特征。对于每种情况，在特征选择过程后，保留具有良好可预测性的特征；剩下的特征用23种算法逐一构建模型。采用5重法交叉验证模型，重复5次。选择准确率最高的算法模型作为SICH血肿增大（HE）的预测模型，以其平均参数AUC、准确率、灵敏度、特异性、F1评分、阳性预测值（PPV）、阴性预测值（NPV）、假阳性率（FPR）、假阴性率（FNR）、假发现率（FDR）作为评价指标。结果：每个脑血肿共有1227个纹理特征。经过特征选择后，剩下4个特征（wavelet-LHL mean、wavelet-LLL _ Idm、wavelet-LLL _跑长非均匀化和wavelet-LLL _contrast）构建预测模型。在23种模型算法中，线性支持向量分类器的准确率最高（72.6%），最终被选为预测模型，其AUC、准确率、灵敏度、特异性、F1评分、PPV、NPV、FPR、FNR、FDR分别为0.729、0.726、0.717、0.736、0.714、0.736、0.741、0.264、0.283、0.264。结论：基线NCCT图像上脑血肿的放射组学特征对预测SICH的HE有很好的效果。
Objective: To explore the value of radiomics features on non-contrast computed tomography (NCCT) in predicting early enlargement of spontaneous intracerebral hemorrhage (SICH).Patients and Methods: 167 patients with SICH were divided into enlarged hematoma and non-enlarged hematoma groups based on the volume of hematoma on 24-h follow-up CT images > 30% and/or 6 ml of the baseline NCCT. The baseline NCCT images of all cases were imported into radiomics software to extract the radiomics features of the initial hematoma. For each case, the features with good predictability were retained after the feature-selected process; the remaining features were used to construct model with 23 algorithms one-by-one. A 5-fold method was used to cross-validate the model and repeated 5 times. The algorithm model with the highest accuracy was selected as predictive model for hematoma enlargement (HE) in SICH, its average parameters including AUC, accuracy, sensitivity, specificity, F1 score, positive predictive value (PPV), negative predictive value (NPV), false positive rate (FPR), false negative rate (FNR),and false discovery rate (FDR) were taken as evaluating indicators.Results: A total of 1227 texture features of each cerebral hematoma were obtained. After the feature-selected process, 4 features (wavelet-LHL mean, wavelet-LLL _ Idm, wavelet-LLL _run length non-uniformity normalized, and wavelet-LLL _contrast) remained to construct the predictive models. Among 23 model algorithms, Linear Support Vector Classifier showed the highest accuracy (72.6%), and eventually was selected as the predictive model, its AUC, accuracy, sensitivity, specificity, F1 score, PPV, NPV, FPR, FNR, and FDR were 0.729, 0.726,0.717,0.736,0.714, 0.736, 0.741, 0.264, 0.283 and 0.264, respectively.Conclusion: Radiomics features of cerebral hematoma on baseline NCCT images showed good performance in predicting HE of SICH.