超声纹理分析用于预测急性心肌梗死
研究纳入684名参与者,基于2腔和4腔心尖切面超声图像提取手工与深度学习 ultrasomics 特征,训练机器学习模型识别梗死心肌;机器学习模型识别心肌梗死的AUC为0.87,高于迁移学习深度特征的0.74(P<0.0001),且在存在心肌功能障碍时仍有诊断价值。
要点
- 机器学习 ultrasomics 模型识别心肌梗死的AUC为0.87,高于迁移学习深度特征的0.74(P<0.0001)。
- 在平均纵向应变<16%的心肌功能障碍者中,机器学习概率鉴别急性心肌梗死的AUC为0.84。
- 平均纵向应变联合机器学习概率的AUC为0.86,高于单独纵向应变的0.80(P=0.02)。
核心结论
- 心脏 ultrasomics 有潜力区分健康与梗死心肌,原文指出需在多样人群中验证以明确其超越常规超声心动图在心肌组织表征中的作用与增量价值。
研究设计与样本
- 研究纳入684名参与者,数据来源包括回顾性单中心匹配病例对照数据集、前瞻性多中心匹配临床试验数据集和开源国际多厂商数据集;从2腔和4腔超声心动图切面提取手工和基于深度迁移学习的 ultrasomics 特征,采用留一来源交叉验证进行外部验证训练机器学习模型。
关键影像表现/征象
- 利用 Manhattan 图可视化 ultrasomics 特征可区分梗死与非梗死节段(P<0.001),并有助于梗死心肌的参数化可视化。
诊断标准与关键数值
- 机器学习模型识别心肌梗死的AUC为0.87(95%CI:0.84-0.89),迁移学习深度特征AUC为0.74(95%CI:0.70-0.77),P<0.0001。
- 调整常规超声心动图参数后,机器学习概率仍为心肌梗死的独立预测因子(校正OR:1.03,95%CI:1.01-1.05;P<0.0001)。
- 在平均纵向应变<16%的心肌功能障碍者中,机器学习概率鉴别急性心肌梗死的AUC为0.84(95%CI:0.77-0.89)。
- 平均纵向应变联合机器学习概率的AUC为0.86(95%CI:0.80-0.91),高于单独平均纵向应变的AUC 0.80(95%CI:0.72-0.87),P=0.02。
临床应用与阅片要点
- 原文结论指出心脏 ultrasomics 有潜力区分健康与梗死心肌,并强调需在多样人群中验证以明确其作用与增量价值。
局限性
- 原文指出需在多样人群中验证以明确其作用与超越常规超声心动图在心肌组织表征中的增量价值。
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
BACKGROUND: Acute myocardial infarction (MI) alters cardiomyocyte geometry and architecture, leading to changes in the acoustic properties of the myocardium.
OBJECTIVES: This study examines ultrasomics-a novel cardiac ultrasound-based radiomics technique to extract high-throughput pixel-level information from images-for identifying ultrasonic texture and morphologic changes associated with infarcted myocardium.
METHODS: The authors included 684 participants from multisource data: a) a retrospective single-center matched case-control dataset; b) a prospective multicenter matched clinical trial dataset; and c) an open-source international and multivendor dataset. Handcrafted and deep transfer learning-based ultrasomics features from 2- and 4-chamber echocardiographic views were used to train machine learning (ML) models with the use of leave-one-source-out cross-validation for external validation.
RESULTS: The ML model showed a higher AUC than transfer learning-based deep features in identifying MI (AUC: 0.87 [95% CI: 0.84-0.89] vs AUC: 0.74 [95% CI: 0.70-0.77]; P < 0.0001). ML probability was an independent predictor of MI even after adjusting for conventional echocardiographic parameters (adjusted OR: 1.03 [95% CI: 1.01-1.05]; P < 0.0001). ML probability showed diagnostic value in differentiating acute MI, even in the presence of myocardial dysfunction (averaged longitudinal strain [LS] <16%) (AUC: 0.84 [95% CI: 0.77-0.89]). In addition, combining averaged LS with ML probability significantly improved predictive performance compared with LS alone (AUC: 0.86 [95% CI: 0.80-0.91] vs AUC: 0.80 [95% CI: 0.72-0.87]; P = 0.02). Visualization of ultrasomics features with the use of a Manhattan plot discriminated infarcted and noninfarcted segments (P < 0.001) and facilitated parametric visualization of infarcted myocardium.
CONCLUSIONS: This study demonstrates the potential of cardiac ultrasomics to distinguish healthy from infarcted myocardium and highlights the need for validation in diverse populations to define its role and incremental value in myocardial tissue characterization beyond conventional echocardiography.
引用
- Jamthikar AD, Hathaway QA, Maganti K, et al. Ultrasonic Texture Analysis for Predicting Acute Myocardial Infarction. JACC Cardiovasc Imaging. 2025. PMID: 40815256. doi:10.1016/j.jcmg.2025.06.018
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。