Automated detection of blood clot formation during aneurysm treatment
Neuro-interventional procedures, such as endovascular coiling and flow diversion, have a critical complication risk: thrombus blood clot formation during the procedure. Undetected blood clots can lead to cerebral infarction or death. Current detection methods rely heavily on clinicians' interpretation of imaging, which makes it easy to miss small or rapidly forming blood clots. Radboudumc (Nijmegen), TU/e (Eindhoven), Erasmus MC (Rotterdam), and Hospital Clinic (Barcelona) will work together with Philips (Eindhoven) to develop automated thrombus detection system to enhance procedural safety.
The technology that will be developed within this consortium will increase patient safety and lifts a burden from our healthcare systems due to reduced hospital stays, faster rehabilitation, and less long-term care. The average medical costs associated with cerebral infarction are approximately €11,000 in the first year. An automated thrombus detection system that prevents even a small percentage of such complications could lead to significant cost savings.
By developing a real-time, AI-driven blood clot detection system, the project aims to enhance procedural safety, reduce complications, and improve patient outcomes. Faster detection enables timely intervention, potentially preventing severe neurological damage and reducing recovery times. This directly benefits patients, families, and caregivers, while also lowering healthcare costs.
First, we will use angiographic data from neuro-interventional procedures to develop a real-time, automated blood-clot detection algorithm. The focus is on designing an AI-driven system that accurately detects blood clots, including small or subtle ones, which are often missed by the human eye. This algorithm will form the foundation for future clinical integration. The Open Innovation Platform of Philips will be subsequently used to further refine and evaluate the thrombus detection algorithm. Maturing the technology through offline user testing, system integration into angiography platforms, and systematic evaluation will be important to get the technology closer to market release.
