Development and validation of a digital biomarker for peripheral artery disease (www.nature.com)

Researchers have developed and validated a promising digital biomarker for diagnosing peripheral artery disease (PAD) using photoplethysmography (PPG), a non-invasive method for measuring peripheral blood flow. Analyzing data from over 5,200 legs across 2,362 patients, the study identified significant correlations between various PPG features and the ankle-brachial index (ABI), a standard diagnostic tool for PAD. They built two machine learning models: one based solely on PPG features with an AUC of 0.83, and a second enhanced model that includes clinical information, achieving an AUC of 0.85. Notably, the model demonstrates strong generalizability across different demographics and comorbidities. This development is significant for the AI and ML community, as it underscores the potential of leveraging easily obtainable physiological data for robust medical diagnostics. By establishing a digital biomarker rooted in physiological characteristics, the study paves the way for future research that could integrate the model into clinical workflows, potentially revolutionizing the approach to PAD diagnosis and allowing for earlier detection and intervention in at-risk populations.