Leiden University Medical Center (LUMC), Leiden University and Philips have come together to set up an artificial intelligence (AI) laboratory to accelerate and improve Magnetic Resonance Imaging (MRI) scans. To set up this lab and make MRI more accessible, they have received over €2M from NWO.
The three partners are working together to administer the AI4MRI lab, which operates within the nationwide network of Innovation Centre for Artificial Intelligence Labs, or ICAI labs.
The University of Amsterdam-led ROBUST consortium received a total of €87M from NWO, €2.1M of which will go to AI4MRI. The lab is a part of the SAILS programme, which aims to improve Leiden’s AI ecosystem.
Aim of this collaboration
Currently, MRI exams require patients to lie motionless for 15 minutes or longer in a narrow, noisy space. It’s not a comfortable situation. Additionally, if the patient moves slightly, the MRI image will be somewhat out of focus, making it challenging to detect abnormalities.
Since 2019, researchers from the LUMC, Leiden University, and Philips have successfully collaborated to use AI to speed up MRI scans. The NWO funding will accelerate their research.
With this collaboration, the researchers aim to capture a whole MRI test in under five minutes. To do this, they are using AI, which will require less data to produce a decent MRI image. This could speed up scans considerably.
The team also wants to develop AI that makes MRI scans resistant to patient movement. Additionally, they are concentrating particularly on the heart. A scan is more difficult since this organ is always moving.
The researchers believe that AI will have a significant influence on scan time, scan quality, and ultimately, patient comfort. Besides the patients, this development will also benefit radiology departments, who will be able to operate more effectively.
Within the project, researchers are also looking into how the workplace will change as part of the experiment. For instance, the usage and acceptability of AI MRI scanning can be increased by examining how this technology affects the professionals who use it and the organisational and policy environment in which it is used.
Researchers on this project
The scientific management of the AI4MRI ICAI lab includes LUMC professors Marius Staring and Thijs van Osch from the Department of Radiology; Dr Nicola Pezzotti, Senior Scientist at Philips and Assistant Professor at TU/e; and professor Bram Klievink from Leiden University.
Staring says, “This NWO grant contributes tremendously to strengthening AI research on MRI. And solidifies the good ties between Leiden and Philips,” adds Pezzotti.
Relevant results
Richard Kemkers, Innovation Program Manager at Philips, says, “By working intensively together, we strengthen each other’s knowledge and competence, resulting in new innovative products that can be used worldwide. The first results have already been introduced (SmartSpeed MRI scans), which gives us great confidence for future improvements that are much needed in healthcare.”
In comparison to other MRI modalities, Philips says that the addition of SmartSpeed to the company’s Compressed SENSE MR acceleration engine offers a threefold reduction in MRI scanning time and increases image resolution up to 65 per cent. SmartSpeed (Philips) has already garnered FDA 510(k) clearance.
