Jump to content

News from Fachhochschule Dortmund

Search

Language

Biomedical Informatics

Better blood pressure determination

Published

Fabienne Sahl is the first Master's graduate of the Biomedical Information Technology study program.

Fabienne Sahl has developed a new, better method of measuring blood pressure, earning her the first Master's degree in the young Biomedical Information Technology study program.

First of all, a correction: "blood pressure measurement" is actually the wrong term, as the blood pressure would have to be measured directly inside the artery. However, this only happens during operations and in intensive care units. All other methods that take blood pressure from the outside are therefore correctly called "estimation".

This also applies to the best-known method with the inflatable cuff around the upper arm. This has one major disadvantage: the air pressure required for the estimation compresses the artery so that several minutes must elapse between individual estimations to allow the vessels to expand back to their normal size. It is therefore not possible to determine a blood pressure curve, i.e. the complete temporal representation of the pulse-driven blood circulation.

However, because this curve would contain a lot of important information about the cardiovascular system and the state of health of the person being examined, a simple and reliable method for estimating it would be very helpful for doctors.

Your task: Find the formula

This is where Fabienne Sahl's research idea comes in. Three sensors on the upper arm serve as signal receivers. These measure the pressure that spreads outwards through the tissue when the pulse expands the artery. One of Sahl's tasks was to find a formula (mathematically: a "transfer function") that could be used to convert the signals received into a medically useful blood pressure curve.

Fabienne Sahl used a "deep neural network" of artificial intelligence, or AI for short. Put simply, she fed the AI with the data from the pressure sensors. The AI then adapted the transfer function so that it could in turn estimate the blood pressure based on the pressure signals. At the same time, Fabienne Sahl determined the blood pressure using the already recognized, reliable "volume clamp method". She then compared the blood pressure estimate with the volume-clamp values and readjusted the way the AI works - so that the transfer function can make blood pressure estimates that match the volume-clamp values.

Many advantages and one obstacle

The search for a simple, reliable method for estimating the blood pressure curve is occupying many researchers worldwide. A number of promising methods have been found in recent years, including some AI-supported ones, but they all have one or other disadvantage: for example, excessive data hunger or non-transparent calculation methods.

Fabienne Sahl's method has the charm of being significantly less affected by these disadvantages. In addition, the transfer function allows conclusions to be drawn about the individual composition of the upper arm tissue, which is valuable from a medical point of view.

The method is a promising approach - and the result of a successful master's thesis - but is not yet fully developed. The sensors, whose data is very susceptible to interference, proved to be a problem. But that could soon change: In her doctorate at Fachhochschule Dortmund, Fabienne Sahl would like to work out the solutions resulting from her master's thesis.

Reference: The volume clamp method

In her work, Fabienne Sahl used the volume clamp method as a reference. This works via a blood pressure cuff on the finger that fills with air. Unlike the upper arm cuff, the air pressure adapts to the pulse rate so that it precisely counteracts the fluctuations in the artery and the pressure between the cuff and the artery remains constant throughout. However, even this method, which is considered the most accurate of the non-invasive variants, has disadvantages: for example, it does not allow long-term monitoring because it also compresses the arteries.

The examiners of the Master's thesis were Prof. Dr. Benjamin Menküc and Prof. Dr. Sebastian Zaunseder.