24/08/2026
Pharma Focus Issue 3, Volume 3
AI-Derived Vascular Age: A Deep Learning Approach to Cardiovascular Risk Prediction Among the developments in cardiovascular research, the creation of the AI-VascularAge (AI-V A) model is a very important improvement for identifying the risk of cardiovascular disease early on. The usual way to check stiffness is by determining carotid-femoral pulse wave velocity (cfPWV). This method requires tools and trained people, which makes it hard to use in many places. The AI-V A model solves this problem by using a neural network. It is trained on
pressure waveforms that are not calibrated and are taken in a non-invasive way through brachial,
radial, and carotid tonometry. These waveforms are used to predict the inverse of cfPWV.
Validated in the Framingham Heart Study, AI-V A showed separate links with new cases of cardiovascular disease (hazard ratio 1.50 per standard deviation), coronary heart disease (HR 1.64), and heart failure (HR 1.65) even after taking into account traditional risk factors. The main advantage of AI-V A is that it takes features from waveforms that do not depend on calibrated pressure amplitude, heart rate, or body size, which helps reduce measurement bias while allowing assessment at the point of care.
This new development is important because it makes a complicated medical test easier to use. It can help us find people who are at risk of heart problems in an earlier stage. This can potentially prevent progression of cardiovascular problems in the first place. The test can do this by showing us if the blood vessels are getting older faster than they should be. However, this research was based on experiments done on people of European descent. More clinical trials should be performed on people of different races to establish its effectiveness.
Source- https://pmc.ncbi.nlm.nih.gov/articles/PMC12960199/
Writeup and poster by
Sumaia Sayed Hridita
ID-23146001
