Abstract: Objective  To construct and validate a nomogram prediction model based on the modified total burden score of cerebral small vessel disease for the precise prediction of spontaneous hemorrhagic transformation in acute ischemic stroke patients without reperfusion therapy.
Methods   Acute ischemic stroke patients without reperfusion therapy hospitalized in Sixth People’s Hospital of Nanhai District, Foshan from January 2021 to July 2024 were prospectively enrolled. They were randomly divided into a training set and an internal validation set at a ratio of 7∶3. Additionally, patients who met the inclusion and exclusion criteria and were admitted to the same hospital during different time periods (January to December 2020, and August 2024 to October 2025) were recruited as an external validation set. Data including gender, age, past medical history, hematological and imaging examination results were collected for all patients. Imaging markers of cerebral small vessel disease were evaluated based on cranial MRI examination results, and each marker was assigned a value according to its location and severity to calculate the modified total burden score of cerebral small vessel disease. The diagnosis of spontaneous hemorrhagic transformation was determined based on the results of the second cranial CT or MRI examination during hospitalization. With the occurrence of spontaneous hemorrhagic transformation as the dependent variable, univariate and multivariate logistic regression analyses were performed on the training set to screen for predictive factors, which were then used to construct the nomogram prediction model. Calibration curve was used to assess the consistency of the model, ROC curve was applied to evaluate the prediction efficacy of the model, and decision curve analysis and clinical impact curve were used to evaluate the clinical application value of the model.
Results  A total of 1430 acute ischemic stroke patients without reperfusion therapy were included, with a mean age of (67.9±10.4) years, including 716 females (50.1%). The training set comprised 547 patients, with a mean age of (68.2±10.4) years, including 279 females (51.0%). Multivariate logistic regression analysis showed that the modified total burden score of cerebral small vascular disease (OR 2.817, 95%CI 2.210-3.591, P<0.001) and large hemispheric infarction (OR 2.642, 95%CI 1.115-6.260, P=0.027) were independent risk factors for spontaneous hemorrhagic transformation after acute ischemic stroke. The calibration curve of the nomogram prediction model in the training set showed good agreement between the predicted and observed values. The ROC curve showed an AUC of 0.835 (95%CI 0.789-0.880), indicating that the model had good predictive efficacy. The decision curve analysis results revealed that the net benefit was the highest when the threshold probability was from 0.06 to 0.77. The clinical impact curve analysis suggested that the model had an acceptable cost-benefit ratio, indicating high clinical application value. The internal validation set included 235 patients, with a mean age of (68.3±10.4) years, including 119 females (50.6%), and the ROC curve showed an AUC of 0.847 (95%CI 0.785-0.910). The external validation set included 648 patients, with a mean age of (67.4±10.4) years, including 318 females (49.1%), and the ROC curve showed an AUC of 0.870 (95%CI 0.795-0.931).
Conclusions  The nomogram prediction model constructed in this study can effectively predict the risk of spontaneous hemorrhagic transformation after acute ischemic stroke.

Key words: Modified total burden score of cerebral small vessel disease; Acute ischemic stroke; Spontaneous hemorrhagic transformation; Nomogram

摘要： 目的 构建并验证基于改良脑小血管病总负荷评分精准预测非再灌注治疗急性缺血性卒中患者自发性出血转化的列线图预测模型。
方法 前瞻性纳入2021年1月—2024年7月于佛山市南海区第六人民医院住院的非再灌注治疗急性缺血性卒中患者，按7∶3比例随机分为训练集与内部验证集。另纳入该院不同时间段（2020年1—12月及2024年8月—2025年10月）符合纳入排除标准的非再灌注治疗急性缺血性卒中患者作为外部验证集。收集患者的性别、年龄、既往病史、血液学及影像学检查资料。根据头颅MRI检查结果评估不同脑小血管病影像学标志物，并依据不同部位及严重程度进行赋值，计算改良脑小血管病总负荷评分。自发性出血转化的判定依据是住院期间第2次头颅CT或MRI检查结果。以是否发生自发性出血转化为因变量，在训练集中进行单因素及多因素logistic回归分析筛选预测因素，进而构建列线图预测模型。采用校准曲线评估模型的一致性，应用ROC曲线评估模型的预测效能，使用决策曲线分析和临床影响曲线评估模型的临床应用价值。
结果 共纳入1430例非再灌注治疗急性缺血性卒中患者，平均年龄为（67.9±10.4）岁，女性716例（50.1%）。训练集共547例，平均年龄为（68.2±10.4）岁，女性279例（51.0%），多因素logistic回归分析显示，改良脑小血管病总负荷评分（OR 2.817，95%CI 2.210～3.591，P<0.001）、大面积梗死（OR 2.642，95%CI 1.115～6.260，P=0.027）是急性缺血性卒中后自发性出血转化的独立危险因素。该列线图预测模型的训练集校准曲线显示，预测值和观测值吻合良好；ROC曲线显示AUC为0.835（95%CI 0.789～0.880），提示模型具有良好的预测效能；决策曲线分析结果显示，当阈值概率在0.06～0.77时净收益最高；临床影响曲线分析结果提示，模型具有可接受的成本效益比，表明其具有较高的临床应用价值。内部验证集共235例，平均年龄为（68.3±10.4）岁，女性119例（50.6%），ROC曲线显示AUC为0.847（95%CI 0.785～0.910）。外部验证集共648例，平均年龄为（67.4±10.4）岁，女性318例（49.1%），ROC曲线显示AUC为0.870（95%CI 0.795～0.931）。
结论 本研究构建的列线图预测模型可有效预测急性缺血性卒中后自发性出血转化的发生风险。

关键词: 改良脑小血管病总负荷评分; 急性缺血性卒中; 自发性出血转化; 列线图