From 911 to rehabilitation
For one of the two labs, Erasmus MC, in collaboration with the EUR, will enter into a new partnership with Philips Healthcare. Together they will improve the patient journeys of patients suffering from a stroke. ‘A stroke is an acute problem,’ says Niessen, ‘but it has a long aftermath. We want to use artificial intelligence to help the patient from A to Z. That starts with dialling 911 and ends with rehabilitation.’
From the very first moment during the patient journey, choices have to be made. Niessen: ‘The patient enters the hospital. You need to diagnose them correctly with imaging by a CT scan. After that, you want to offer the patient the best treatment and ultimately the best rehabilitation.’
Over the next ten years, a total of 10 PhD students will work in the laboratory with Philips to make those choices easier. ‘They will create an environment in which data collection is possible along the patient’s entire timeline. They then can use AI to extract information from this data. For example, which type of patient is better suited for which treatment. We can eventually use that information to provide personalized treatment to improve patient outcomes.’
For now, this lab is starting with the patient journey for stroke. Eventually, this initiative will include other patient journeys by joining Convergence, a collaboration between the EUR, TU Delft and Erasmus MC.
Neuro-radiologist Aad van der Lugt will be the driving force behind the second lab. In doing so, Erasmus MC continues its long-standing collaboration with General Electric Healthcare. An MRI examination must and can be better geared to the individual patient,’ he says. What would that look like, according to him?
Van der Lugt envisions a faster and more effective scanner: ‘An MRI scanner that determines the protocols in advance by providing information about the patient. One that can adjust itself during the scan using information from the first series of scans. Or one that doesn’t have to start over if the patient moves during the scan, but can correct motion artefacts automatically.’
Van der Lugt also hopes that AI can eventually help radiologists interpret the scans. Algorithms could, for example, find and analyze small details that are not or barely visible. The Radiology Lab will be home to 10 PhD students over the next 10 years.
The ROBUST consortium will establish a total of 17 ICAI labs, led by the principal applicant and UvA professor Maarten de Rijke. NWO is providing 25 million euros, and participating companies and UMCs will invest the remaining sum. ‘It’s a huge investment in AI,’ says Niessen. ‘That’s because the impact of AI in healthcare and society will be enormous, provided we can rely on it properly. We need to make sure that this powerful technology is used responsibly in daily clinical practice. That’s what we’ll be working hard on over the next 10 years.’