文献速递外周血管胸部US

颈动脉斑块影像组学特征联合机器学习模型预测心血管事件

Radiomic Carotid Plaque Features Integrated into Machine Learning Models for Cardiovascular Risk Prediction.
PubMed 文献 · Ultrasound Med Biol · 2026-06-12

本研究纳入493例患者,通过颈动脉超声获取斑块图像,结合临床资料、手动FOVUS测量及半自动影像组学特征,构建机器学习模型预测5年内主要不良心血管事件(MACE)。结果显示,临床+FOVUS+影像组学联合模型预测准确率最高(0.958±0.023),优于单一数据源模型。研究表明,自动及半自动斑块定量分析均可为心血管风险分层提供预测价值。

要点

  • 联合临床、FOVUS及影像组学数据的ML模型预测MACE准确率最高达0.958±0.023
  • 影像组学单独预测准确率较低(0.705±0.036),需结合临床与超声测量数据
  • 半自动斑块定量分析可作为手动测量的补充,二者均能生成有效预测特征

核心结论

  • 临床资料、FOVUS测量与影像组学特征联合的机器学习模型可有效区分MACE与非MACE患者,预测准确率显著优于随机(p=2×10⁻⁷)
  • 自动与半自动斑块定量分析均能生成预测性影像组学特征,可用于心血管风险分层

研究设计与样本

  • 前瞻性队列研究,493例颈动脉超声患者,随访5年通过病历审查收集MACE结局
  • 单中心研究,超声扫描获取颈动脉斑块影像,采用半自动影像组学特征提取
  • 特征选择采用ReliefF算法,样本与特征比10:1,最终纳入11个特征
  • 四种机器学习分类器,使用10折交叉验证评估模型性能

关键影像表现/征象

  • 5年随访中144例(29%)发生MACE,终点事件包括死亡、不稳定心绞痛、心肌梗死、卒中、TIA、颈动脉内膜剥脱术、冠脉搭桥、经皮冠脉介入
  • 影像组学特征需结合临床和FOVUS数据才能显著提升预测性能,单独使用时准确率较低
  • FOVUS与影像组学联合模型准确率(0.935±0.043)高于单独任一种数据,提示两者信息互补

诊断标准与关键数值

  • 最优模型(x-gradient boost)预测准确率:临床+FOVUS+影像组学=0.958±0.023
  • 各模型准确率:临床单独=0.849±0.039,FOVUS单独=0.875±0.029,影像组学单独=0.705±0.036
  • 临床+FOVUS联合=0.885±0.023,临床+影像组学联合=0.804±0.053,FOVUS+影像组学联合=0.935±0.043
  • MACE终点定义:死亡、不稳定心绞痛、心肌梗死、卒中、短暂性脑缺血发作、颈动脉内膜剥脱术、冠状动脉旁路移植术、经皮冠状动脉介入
  • 模型性能检验采用10折交叉验证,p=2×10⁻⁷(最优模型优于随机水平)

临床应用与阅片要点

  • 日常报告中可同时记录斑块的超声形态学特征(如厚度、回声)与影像组学定量参数,为风险分层提供补充信息
  • 影像组学单独预测效能有限,应作为传统临床和超声测量指标的辅助手段,不宜单独用于评估
  • 联合模型建议整合临床危险因素、FOVUS测量和影像组学特征,以最大化MACE预测能力
  • 需注意本研究中FOVUS测量和影像组学特征均为颈动脉斑块特异性指标,其他血管床斑块可能不适用

局限性

  • 单中心研究,样本来自一个机构,泛化性受限
  • MACE结局通过5年病历审查获取,可能受记录完整性影响
  • 影像组学特征提取为半自动方式,依赖操作者初始标注,存在一定主观性

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

OBJECTIVE: Carotid plaque detected by ultrasound is associated with major adverse cardiovascular events (MACE) and can be characterized using manual or automated radiomic analysis. The generated plaque characteristics may have complex interdependencies for which machine learning (ML) may provide predictive modeling. The objective of the study was to develop a ML model to predict MACEs from clinical variables, manual focused vascular ultrasound (FOVUS) measurements, and semi-automated radiomic ultrasound features.

METHODS: Carotid ultrasound scans were performed on 493 patients and MACE outcomes were collected over 5 y by medical chart review. ML-based models were compared that incorporated clinical characteristics, manual FOVUS measurements, and quantitative radiomic features. Feature selection was performed via ReliefF, with a sample:feature ratio of 10:1 yielding 11 top features for training the ML model. Four ML classifiers were developed using 10 K-fold cross-validation.

RESULTS: Over 5 years, 144 patients (29%) experienced a MACE outcome (death, unstable angina, myocardial infarction, stroke, transient ischemic attack, carotid endarterectomy, coronary artery bypass graft, or percutaneous coronary intervention). The best model (x-gradient boost) performed significantly better than chance alone (p = 2 × 10-7), with the highest average prediction accuracies (clinical data only: 0.849 ± 0.039, FOVUS only: 0.875 ± 0.029, radiomics only: 0.705 ± 0.036, clinical and FOVUS data: 0.885 ± 0.023, clinical and radiomics data: 0.804 ± 0.053, FOVUS and radiomic data: 0.935 ± 0.043, all data: 0.958 ± 0.023).

CONCLUSION: The combination of clinical, FOVUS and radiomics data provides strong predictive power for distinguishing MACE from non-MACE. Moreover, both automated and manual plaque quantification can generate predictive features.

来源:Ultrasound Med Biol,查看原文。版权归原作者所有。

引用

  1. Hu, Mojtahedi, Duli, et al. Radiomic Carotid Plaque Features Integrated into Machine Learning Models for Cardiovascular Risk Prediction. Ultrasound Med Biol. 2026. PMID: 42285860. doi:10.1016/j.ultrasmedbio.2026.05.006

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。