At the Medica international trade fair for the medical sector in Düsseldorf, Verhaert proposed a new AI-based technology that directly measures blood pressure through photoplethysmographic (PPG) signals in fingers, wrists and upper arms. It can be used for long-term monitoring of hypertension patients without the need of a sphygmomanometer. This enables new applications such as personalized drug dosage and dosage adaptation based on exact situationally established blood pressure measurements.
The technology combines feature-based deep learning and featureless machine learning algorithms, which results in a mean of 7 mm Hg of absolute error on the diastole and systolic blood pressure. Many people could benefit from this technology, as the European Society of Cardiology states that, even in lower-risk cardiovascular disease populations, such as Belgium and Germany, many people risk dying of heart diseases within the next 10 years.
“Verhaert’s AI team has worked with KU Leuven and UZ Leuven with the support of the Flemish Agency for Innovation and Entrepreneurship – Vlaio – over a number of years to develop and clinically validate the technology. Medical practitioners can now access data of patients more easily and in more natural situations. It allows improved monitoring of diseases such as hypertension and diabetes in a more meaningful way,” says Jochem Grietens, head of the Verhaert AILab.