Scholarlink

<p>Peripheral artery disease (PAD) is a common manifestation of atherosclerotic cardiovascular disease (ASCVD) that is underdiagnosed in clinical practice. Photoplethysmography (PPG) serves as a widely available tool that captures peripheral vascular physiology, yet the quantitative links between PPG signal characteristics and the presence of PAD are underexplored. In analyzing 5,237 legs from <i>N</i> = 2362 unique patients, we find significant correlations with multiple PPG features and the ankle-brachial index (ABI), a commonly used non-invasive diagnostic test for PAD. Using these explainable features, we develop a machine learning model to detect PAD solely from PPG features (AUC = 0.83) and develop an enhanced model incorporating clinical information (AUC = 0.85). Additionally, our model is highly generalizable, performing similarly across demographics and comorbidities. These findings represent an initial step toward identifying an accessible, physiologically grounded digital biomarker associated with PAD, and lay the foundation for prospective studies to evaluate performance across clinical workflows and reference standards.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development and validation of a digital biomarker for peripheral artery disease

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月12日

查看全文
  • Mattheus Ramsis,
  • Ava J. Fascetti,
  • Mustafa H. Naguib,
  • Shamim Nemati,
  • Pam R. Taub,
  • Christopher A. Longhurst,
  • Elsie G. Ross,
  • Edward J. Wang

摘要

Peripheral artery disease (PAD) is a common manifestation of atherosclerotic cardiovascular disease (ASCVD) that is underdiagnosed in clinical practice. Photoplethysmography (PPG) serves as a widely available tool that captures peripheral vascular physiology, yet the quantitative links between PPG signal characteristics and the presence of PAD are underexplored. In analyzing 5,237 legs from N = 2362 unique patients, we find significant correlations with multiple PPG features and the ankle-brachial index (ABI), a commonly used non-invasive diagnostic test for PAD. Using these explainable features, we develop a machine learning model to detect PAD solely from PPG features (AUC = 0.83) and develop an enhanced model incorporating clinical information (AUC = 0.85). Additionally, our model is highly generalizable, performing similarly across demographics and comorbidities. These findings represent an initial step toward identifying an accessible, physiologically grounded digital biomarker associated with PAD, and lay the foundation for prospective studies to evaluate performance across clinical workflows and reference standards.