In the last decade, research in developing systems for monitoring clinical parameters has grown exponentially, driven by the increasing demand for continuous and remote observation. Monitoring cardiovascular diseases (CVDs) is crucial, as they remain the leading cause of mortality worldwide. Pulse Wave Velocity (PWV) is a key clinical indicator for assessing arterial stiffness, highly correlated with CVDs. Current methods for measuring PWV require acquiring pulse wave signals at the carotid and femoral sites, configurations that are widely validated in the literature. However, accessing these specific anatomical points presents a significant challenge for developing wearable devices for continuous or home-based PWV monitoring. The present study proposes an innovative approach using a regression algorithm to predict the carotid-femoral PWV value from pulse waves acquired at more accessible sites, including the carotid, radial, and tibial, compatible with wearable use. The study was conducted on a cohort of 90 voluntary healthy participants at the Candiolo Cancer Institute FPO-IRCCS. The best-performing carotid-radial models show a mean error (ME) of less than 0.1 m/s, with a standard deviation error (SDE) below 1 m/s. Similarly, the radial-tibial models exhibit a mean error of the same magnitude but with an even smaller standard deviation error, below 0.9 m/s. These results highlight the strong predictive capabilities of the developed models, offering promising prospects for future clinical implementation aimed at non-invasive and potentially continuous vascular health monitoring.
Pogliano et al. (Mon,) studied this question.