Date added: 2025-12-22
Artificial intelligence increasing the chances of early detection of brain aneurysms
Data instead of scalpels – a paradigm shift in diagnostics
The project aims to create a tool that will enable the risk of aneurysm and its rupture to be assessed on the basis of laboratory data, medical history and descriptions from medical records. This is a groundbreaking approach – until now, it was believed that such an assessment was only possible on the basis of imaging tests such as tomography or angiography.
Researchers assume three stages of work:
- Predicting the risk of aneurysm rupture,
- Assessing the likelihood of a patient belonging to a risk group for aneurysym occurrence,
- Developing a clinical calculator and application for doctors to support decisions related to the treatment of at-risk patients.
“The theme of our research project is: ‘From data to diagnosis. From diagnosis to saved lives’,” emphasized Patryk Jasik, PhD, Eng., team leader from the Gdańsk University of Technology, and Justyna Fercho, MD, PhD, team leader from the Medical University of Gdańsk and the University Clinical Centre in Gdańsk.
Hundreds of thousands of data points and machine learning
The models developed by scientists analyse routine laboratory test results and textual medical records. Data from over 60,000 patients treated at the University Clinical Centre in Gdańsk between 2006 and 2024 was used to create them.
The first models have already achieved very promising results, with over 77% effectiveness and approximately 80% sensitivity, indicating significant clinical potential.
In the first stage, 26 laboratory test results and predictors based on textual medical documentation were used, including test results such as glucose, platelets, creatinine, sodium, MCH, MPV, lymphocytes, potassium, and risk factors such as nicotine addiction, diabetes, and hypertension. Already at this stage, statistically significant differences were found between patients with ruptured and unruptured aneurysms. However, in the case of patients with ruptured aneurysms, the tests were always based on data collected before the rupture.
“The data was prepared very precisely for the analysis and modelling stages, with particular emphasis on safeguards against data leaks. We used modern predictive models, including TabNet, and ensured their multi-criteria validation, which allowed us to avoid overestimating the results. The accuracy of the selected models exceeds 80 per cent, and in some analyses even reaches over 90 per cent,” explained Dr Jasik. “Using such methods is currently highly desirable by doctors,” noted Dr Fercho. “We can consciously use the knowledge on which the probability of aneurysm risk is based by the model to diagnose patients more effectively,” she added.
Language models support documentation analysis
A new element of the project is the use of large language models (LLMs) such as LLaMA, PLLuM and Bielik. Their task is to automatically extract information from medical descriptions, which significantly increases the effectiveness of predictive models and allows for the analysis of data that was previously difficult to process by computer.
“Using language models allows for very rapid processing and analysis of medical notes, which are a very valuable source of data in the entire diagnosis of aneurysms,” explained the scientist.
Aneurysms – a health tragedy and enormous costs
The problem of aneurysms can affect any of us. Statistically, as many as 1 in 50 adults may have an unruptured brain aneurysm. In Poland, this means hundreds of thousands of people are potentially at risk.
Scientists have calculated the average costs of treating patients with unruptured aneurysms. Currently, there are two main methods used to secure unruptured aneurysms.
“The first is surgical clipping, which is an invasive method requiring the opening of the skull, and its average cost is approximately PLN 30,000,” said Dr Justyna Fercho.
“The second, less invasive endovascular method is embolisation, which involves inserting materials that close the aneurysm through a thin catheter (usually through the femoral artery). Most commonly used are platinum coils, which fill the interior of the aneurysm, causing it to close, cut off from circulation and form a clot, preventing rupture or rebleeding. In the case of wide-neck or complex aneurysms, stent-assisted coiling or flow diverters are additionally used, which are placed in the parent vessel to change the haemodynamics and promote natural closure of the aneurysm. The cost of such an endovascular procedure is approximately PLN 25,000–40,000, depending on the material used.”
Recognised in the I3HIES programme
The project ‘AI-Powered Medical Software for Predicting the Likelihood of Intracranial Aneurysms’ is the result of close cooperation between research teams from the Medical University of Gdańsk, the Gdańsk University of Technology (FAPM and the BioTechMed Center), the University Clinical Centre, and industry partners specialising in data science.
The project has now been selected for the international I3HIES acceleration programme, which supports the most promising innovations in healthcare and helps develop them through to the investment and commercialisation stages.
The I3HIES programme operates under the Interregional Innovation Investments (I3) Instrument financed by the European Regional Development Fund. Its task is to support health projects with high implementation potential, especially those that require advanced validation, overcoming regulatory barriers and preparation for investment.
“Thanks to our participation in this programme, we will be able to check whether our solution is suitable for widespread implementation. Among other things, we will examine how to prepare our product so that European governments can purchase it for their citizens. Above all, this is an opportunity to showcase our research in other countries and popularise this knowledge so that patients from other countries also have access to an innovative diagnostic tool that increases the effectiveness of detecting the potential presence of aneurysms in the brain,” concluded Dr Jasik.