NON-INVASIVE ASSESSMENT OF VASCULAR HEALTH USING PHOTOPLETHYSMOGRAPHIC SENSORS: From In Vitro Arterial Stiffness to In Vivo Vascular Aging Characterization

Photoplethysmography (PPG) is a low-cost and non-invasive optical technique, already widely integrated into consumer wearable devices, that measures blood volume changes in the microvascular bed. Although its clinical potential extends well beyond heart rate and oxygen saturation monitoring, the morphological richness of the PPG signal remains largely underexploited. Arterial stiffness and vascular aging are among the earliest and most significant markers of cardiovascular risk, yet their assessment still relies on specialised clinical techniques unsuitable for continuous, large-scale screening. In this context, PPG offers a promising platform to bring vascular diagnostics closer to everyday health monitoring.This thesis develops an integrated experimental framework for the non-invasive assessment of vascular health using PPG sensors, following a coherent progression from controlled in vitro conditions to clinically heterogeneous populations. The work is articulated in four interconnected studies. First, PPG sensors were employed to estimate arterial wall stiffness through pulse transit time and pulse wave velocity measurements on silicone phantom models, representing different physiological vascular conditions, within a mock circulatory loop. Second, the second derivative of the PPG signal, known as the acceleration plethysmogram (APG), was investigated as a route to vascular aging assessment: a dual in vitro–in vivo design revealed strong correlations between APG-derived morphological indices and both wall stiffness and chronological age, supporting their role as non-invasive surrogates of vascular health. Third, the APG analysis was systematically extended to multiple wavelengths and anatomical sites through a custom-built wearable prototype, identifying the optimal optical configuration and demonstrating compatibility with the dominant smart and wearable devices. Fourth, the discriminative power of APG-derived indices was evaluated on a clinically heterogeneous cohort including both healthy and pathological subjects, demonstrating that these indices capture a vascular aging component that is clinically informative beyond chronological age and can effectively distinguish between healthy and diseased vascular states.Overall, this thesis establishes a unified methodological framework that progresses from physical interpretability to clinical utility, combining in vitro mechanical validation, morphological signal analysis, multi-wavelength optimisation and clinical-group discrimination. The proposed approach lays the groundwork for the integration of PPG-based vascular biomarkers into wearable devices, supporting the transition towards continuous, accessible and large-scale non-invasive cardiovascular monitoring.