多模态临床-MRI影像组学模型对急性缺血性脑卒中患者静脉溶栓后发生早期神经功能恶化的预测价值

孙玉志 ,  喻博 ,  郑健 ,  郭文力

中国医科大学学报 ›› 2026, Vol. 55 ›› Issue (7) : 650 -654.

PDF (520KB)
中国医科大学学报 ›› 2026, Vol. 55 ›› Issue (7) : 650 -654. DOI: 10.12007/j.issn.0258‐4646.2026.07.014
论著

多模态临床-MRI影像组学模型对急性缺血性脑卒中患者静脉溶栓后发生早期神经功能恶化的预测价值

作者信息 +

Predictive value of multimodal clinical-MRI radiomics model for early neurological deterioration after intravenous thrombolysis in patients with acute ischemic stroke

Author information +
文章历史 +
PDF (532K)

摘要

目的 基于多模态临床-磁共振成像(MRI)影像组学联合模型预测急性缺血性脑卒中(AIS)患者静脉溶栓(IVT)后发生早期神经功能恶化(END)的风险。方法 选取2021年1月至2025年6月间我院收治的200例AIS患者,根据是否发生END分为END组(n = 58)和非END组(n = 142),将数据集按照7∶3分为训练集(n = 140)和验证集(n = 60)。基于磁敏感加权成像(SWI)、弥散加权成像(DWI)筛选的影像组学特征和有意义的临床-MRI特征,通过logistic回归分别构建SWI组学模型、DWI组学模型、SWI+DWI组学模型及联合模型。采用受试者操作特征(ROC)曲线、临床决策(DCA)曲线、校准曲线及DeLong检验评估各模型的预测效能。结果 在验证集中,联合模型曲线下面积(AUC)最高,为0.915。SWI+DWI组学模型AUC为0.897,均高于单序列SWI组学模型(AUC=0.814)和DWI组学模型(AUC=0.792),差异均有统计学意义(P < 0.05)。基于临床-MRI特征筛选后得到的4个有意义特征[突出血管征、DWI梗死体积、ADC、溶栓前美国国立卫生院卒中量表(NIHSS)评分]构建临床-MRI模型,其验证集AUC为0.723。DeLong检验结果显示,联合模型的AUC明显高于临床-MRI模型(P < 0.05),与SWI+DWI组学模型AUC比较,差异无统计学意义(P > 0.05)。校准曲线和DCA曲线结果显示,联合模型具有更高的预测精准度和临床净获益率。结论 多模态临床-MRI影像组学联合模型可有效预测AIS患者IVT后发生END的风险,为临床制定个体化预防治疗方案提供依据。

Abstract

Objective To investigate the risk of early neurological deterioration(END)after intravenous thrombolysis(IVT)in acute ischemic stroke(AIS)using multimodal clinical magnetic resonance imaging(MRI)radiomics combined models. Methods A total of 200 patients with AIS admitted to our hospital from January 2021 to June 2025 were selected. Based on the occurrence of END,patients were categorized into an END(n = 58)or non-END(n = 142)group. The dataset was randomly divided in a 7∶3 ratio into a training set(n = 140)and a validation set(n = 60). With the imageomics features screened by magnetic susceptibility weighted imaging(SWI),diffusion weighted imaging(DWI),and relevant clinical-MRI features,logistic regression was used to construct an SWI omics,DWI omics,SWI+DWI omics,and combined models. Model performance was evaluated using receiver operating characteristic(ROC)curves,decision curve analysis(DCA),calibration curves,and the DeLong test. Results In the validation set,the combined model achieved the highest area under the curve(AUC)of 0.915. The AUC of the SWI+DWI omics model was 0.897,which was significantly higher than that of the single-sequence SWI omics(AUC=0.814)and DWI omics models(AUC=0.792,P < 0.05). The clinical-MRI model constructed using four clinically significant features-namely,the asymmetric prominent vein sign,DWI infarct volume,ADC,and the pre-thrombolytic National Institutes of Health Stroke Scale(NIHSS)score-achieved a validation set AUC of 0.723. The DeLong test showed the AUC of the combined model was significantly higher than that of the clinical-MRI model(P < 0.05),whereas no statistically significant difference was observed compared to the SWI+DWI omics model(P > 0.05). Calibration and DCA curves showed that the combined model had superior predictive accuracy and greater clinical net benefit. Conclusion The multimodal clinical-MRI imaging histology-based combined model can effectively predict the risk of post-IVT END in patients with AIS,and provide a basis for the clinical development of individualized preventive treatment plans.

关键词

磁敏感加权成像 / 弥散加权成像 / 影像组学 / 急性缺血性脑卒中 / 早期神经功能恶化

Key words

susceptibility weighted imaging / diffusion weighted imaging / radiomics / acute ischemic stroke / early neurological deterioration

引用本文

引用格式 ▾
孙玉志,喻博,郑健,郭文力. 多模态临床-MRI影像组学模型对急性缺血性脑卒中患者静脉溶栓后发生早期神经功能恶化的预测价值[J]. 中国医科大学学报, 2026, 55(7): 650-654 DOI:10.12007/j.issn.0258‐4646.2026.07.014

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1]

周红俐,杨莎,杨浩,等. 动态CTA结合CTP评估急性缺血性脑卒中患者侧支循环及预后价值[J].中国临床医学影像杂志,2025,36(7):461-464. DOI:10.12117/jccmi.2025.07.002.

[2]

汤笑男,苏利智,余传庆. 急诊绿色通道在急性缺血性脑卒中静脉溶栓治疗中的应用价值[J].中国基层医药,2025,32(8):1244-1248. DOI:10.3760/cma.j.cn341190-20250618-00792.

[3]

卜静英,高铎,袁涛,等. SWI-DWI不匹配在急性缺血性脑卒中病人静脉溶栓预后评估中的价值分析[J].安徽医药,2025,29(7):1359-1364. DOI:10.3969/j.issn.1009-6469.2025.07.018.

[4]

张敏,何顺成,贾鸿运,等. 基于磁共振弥散加权成像的影像组学融合模型预测脑卒中复发[J].中国医学计算机成像杂志,2025,31(4):454-459. DOI:10.19627/j.cnki.cn31-1700/th.2025.04.018.

[5]

中华医学会神经病学分会,中华医学会神经病学分会脑血管病学组. 中国急性缺血性卒中诊治指南2023[J].中华神经科杂志,2024,57(6):523-559. DOI:10.3760/cma.j.cn113694-20240410-00221.

[6]

刘维,苗重昌. DWI联合SWI在急性缺血性脑卒中病情程度评价及溶栓后出血转化预测中的应用价值[J].卒中与神经疾病,2023,30(5):480-485. DOI:10.3969/j.issn.1007-0478.2023.05.009.

[7]

龚燕丽,周莉,姚昕彤,等. CTA联合DWI对急性缺血性脑卒中患者静脉溶栓治疗短期预后评估[J].医学影像学杂志,2024,34(6):142-145. DOI:10.20258/j.cnki.1006-9011.2024.06.035.

[8]

段晓培,韩晓凤,程星遥,等. MRI-DWI与PWI测量参数对老年急性缺血性脑卒中尿激酶溶栓治疗的预后评估价值[J].中国CT和MRI杂志,2023,21(3):14-15,30. DOI:10.3969/j.issn.1672-5131.2023.03.006.

[9]

NAM KW,KWON HM,LEE YS. Infarct growth velocity predicts early neurological outcomes in single subcortical infarction[J].Sci Rep,2023,13:4511. DOI:10.1038/s41598-023-31727-0.

[10]

黄正旺,张水诚,庞杰荣. 磁敏感加权成像与动脉自旋标记成像在脑卒中溶栓后脑出血转化的评估价值[J].中国医学物理学杂志,2024,41(12):1532-1536. DOI:10.3969/j.issn.1005-202X.2024.12.010.

[11]

吕喆,杨欢,王耀辉,等. 基于机器学习的急性缺血性脑卒中静脉溶栓后早期神经功能恶化预测模型的建立与评价[J].中华危重病急救医学,2023,35(9):945-950. DOI:10.3760/cma.j.cn121430-20230601-00413.

[12]

ZHANG LM,WU J,YU RZ,et al. Non-contrast CT radiomics and machine learning for outcomes prediction of patients with acute ischemic stroke receiving conventional treatment[J].Eur J Radiol,2023,165:110959. DOI:10.1016/j.ejrad.2023.110959.

[13]

WEI L,PAN XP,DENG W,et al. Predicting long-term outcomes for acute ischemic stroke using multi-model MRI radiomics and clinical variables[J].Front Med,2024,11:1328073. DOI:10.3389/fmed.2024.1328073.

基金资助

辽宁省教育厅基础研究项目(JYTMS20230098)

AI Summary AI Mindmap
PDF (520KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉