国家神经系统疾病临床医学研究中心数据管理与共享系统
Data Management And Sharing System of China National Clinical Research Center for Neurological Diseases
Predictive etiological classification of acute ischemic stroke through interpretable machine learning algorithms: a multicenter, prospective cohort study
第一作者：siding chen
基本信息
- 负责人：siding chen
- 数据来源： CNSR-Ⅲ
- 论文编号：CNSR-Ⅲ20230113
- 发表时间：2024-09-10
2024-09-10
已发表文章
主要研究背景及目的
背景：不同类型的急性缺血性卒中的预后、复发率和二级预防策略不同，准确预测入院后急性缺血性卒中的病因分型尤为重要。目前，在卒中领域应用机器学习算法预测病因分型的研究极少。机器学习模型可以捕获医疗数据中复杂的非线性关系，并学习其特征来对它们所描述的实体进行分类，这可能适合于模拟病因学评估的过程，故而对已有TOAST/CCS分型进行补充。 目的：本研究旨在利用机器学习和深度学习算法来识别急性缺血性卒中患者病因，并测试模型的区分度和校准度，找到最优模型和最优特征集。对已有病因分型进行补充，建立可以辅助诊断卒中病因工具，帮助临床医生快速对急性缺血性卒中患者进行病因分型。
方法
Methods This study involved patients with the Third China National Stroke Registry (CNSR-III). Nine models, which included Natural Gradient Boosting (NGBoost), Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Light Gradient Boosting Machine (LGBM), Gradient Boosting Decision Tree (GBDT), Adaptive Boosting (AdaBoost), Support Vector Machine (SVM), and logistic regression (LR), were employed to predict large artery atherosclerosis (LAA), small vessel occlusion (SVO), and cardioembolism (CE) using an 80:20 randomly split training and test set. We designed an SFS-XGB with 10-fold cross-validation for feature selection. The primary evaluation metrics for the models included the area under the receiver operating characteristic curve (AUC) for discrimination and the Brier score (or calibration plots) for calibration.
结论
In conclusion, our interpretable ML models, which combine clinical, imaging, and lab data, successfully classify patients with LAA, SVO, and CE, outperforming traditional LR models. Additionally, our model can identify potential CE patients within the SUE group, supplementing existing classifications. This potentially enables clinicians to promptly categorize stroke patients based on their etiologies and initiate optimal prevention and treatment strategies.
主要结果
A total of 5,213 patients were included, comprising 2,471 (47.4%) with LAA, 2,153 (41.3%) with SVO, and 589 (11.3%) with CE. In both LAA and SVO models, the AUC values of the ML models were significantly higher than that of the LR model (P<0.001). The optimal model for predicting SVO (AUC [RF model]=0.932) outperformed the optimal LAA model (AUC [NGB model]=0.917) and the optimal CE model (AUC [LGBM model]=0.846). Each model displayed relatively satisfactory calibration. Further analysis showed that the optimal CE model could identify potential CE patients in the undetermined etiology (SUE) group, accounting for 1,900 out of 4,156 (45.7%).