Algorithms developed by VASCage analyse MRI images of the carotid arteries and brain in search for pathologically critical changes
Warning signals that indicate a potential stroke can be identified using modern imaging techniques like MRI. However, these signals are often hidden within large amounts of data. Artificial intelligence (AI) can assist in detecting critical signs and assessing their severity. The Data Science Department at VASCage, a Centre for Stroke Research in Innsbruck, is developing such algorithms in collaboration with companies, clinics, and universities.
Recognising and evaluating plaques
One of these AI algorithms is designed to detect plaques in the carotid arteries and classify them by risk. “The deposits in the blood vessel walls can be dangerous in various ways. For example, the composition of a plaque plays a crucial role. With high-risk plaques, there’s a strong likelihood that a fragment will break off, travel to the brain via the bloodstream, block a vessel, and cause a stroke. Analysing such plaques can be very time-consuming and complex. That’s why we’re developing an algorithm that automatically identifies plaques in MRI images, precisely segments them, and assesses their risk. This allows treating physicians to take preventive measures if necessary,” explains Marie-Christine Pali, AI developer and mathematician at VASCage.
Support for radiologists
MRI images are three-dimensional and cannot be analysed at a glance. Radiologists must examine each cross-sectional image from top to bottom to identify abnormalities. AI simplifies this tedious process by sifting through the large volumes of image data, thereby easing the workload of medical staff.
VASCage has already developed various algorithms as part of its clinical stroke research. These are used to analyse MRI images of the brain and bone microstructure, among other things, and to generally accelerate the processing of MRI images.
Computers detect more than the human eye
AI is not only a powerful tool for automated image analysis but can also detect details that are invisible to the human eye. While humans can only assess a limited number of parameters, such as gray values, shapes, or structural abnormalities, machine learning can capture characteristics that are not readily visible or even previously unknown. This allows the discovery of novel biomarkers that signal pathological changes. “Image analysis using artificial intelligence is a valuable complement to medical practice. It supports both primary and secondary diagnostics, helping to prevent strokes—whether it’s a first occurrence or a recurrence,” says Karl Fritscher, Head of the Data Science Department at VASCage.
Focus on clinical studies and data sciences
VASCage is a research company specialising in AI-driven image analysis and processing for various medical applications. “In our focus areas—clinical studies and data science—we work with our clients to develop solutions that bring innovations to people quickly and safely,” summarises VASCage CEO Matthias Ullrich.
https://vascage.at/wp-content/uploads/2024/10/PA_KI-zu-Schlaganfall-Risiko.pdf
Picture 1 AI in stroke research (download) photocredit VASCage
Picture 2 AI in stroke research (download) photocredit VASCage
Picture 3 Marie-Christine Pali, AI developer and mathematician at VASCage (download) photocredit VASCage Algorithms developed by VASCage analyse