AI for quantifying MRI: reconstruction & statistical analysis

MRI measures beyond what the eye can see, and artificial intelligence (AI) learns beyond what the human mind can perceive. To quantify this information is what drives us in research and teaching.

We are developing methods throughout the chain of MRI acquisition, reconstruction, quantification and statistical analysis. These studies are deeply motivated by clinical needs, and find their application in different disease types.

Focus

AI for MRI reconstruction and beyond

MRI provides excellent soft tissue contrast, but is limited by extensive imaging times. We have developed the Recurrent Inference Machine (RIM), a physics-informed neural network for accelerating MRI. It has proven successful in learning domain invariant features. We hold a successful track record in participating in reconstruction challenges and have been winning in FastMRI knee and Calgary brain challenges, while being in the top three of the generalization track of the FastMRI Brain challenge.

Clinical decision making increasingly requires further quantification of the imaging data. We are developing end-to-end models for reconstruction and quantification, in parameter mapping and segmentation of structural and functional MRI.

70% of the world population has little or no access to 1.5T and 3T MRI facilities. Low-field MRI scanners find their application in specific clinical settings, are more affordable and require little support in installation and maintenance. Through deep learning techniques, we aim to accelerate the acquisition and efficiently denoise the images.

Together with the Netherlands Cancer Institute, we are developing methods for image guided radiotheraphy.

Funding: TKI-PPP, ABC, ZonMW, industry

Team: Daisy van den Berg, Dimitris Karkalousos, Kai Lønning, Henk Marquering, Nikos Priovoulos, Gustav Strijkers, Matthan Caan

Machine learning for disease characterization in neuroimaging

Machine learning is a powerful tool for early detection, prediction, and treatment of brain diseases in neuroimaging. We focus on methods for robust extraction of quantitative features from medical images. This includes the use of radiomics-based models for disease diagnosis and prognosis.

Large population studies such as UK Biobank provide an opportunity to train machine learning models on large data and account for physiological variation in the data through normative modeling.

In the context of psychiatric disorders, machine learning can be used to predict treatment outcomes and identify patients who are likely to benefit from specific interventions. Machine learning thus has the potential to revolutionize neuroimaging research and clinical practice by providing non-invasive, reliable indicators of brain health, resilience, and vulnerability before clinical manifestations of disease.

Team: Jerke van den Berg, Mingshi Chen, Diogo Fernandes, Maarten Poirot, Liesbeth Reneman, Henk Marquering, Matthan Caan

Funding: TKI-PPP, Eurostars

Output

See more…

Internships

See below for examples of internship projects that we offer. The projects may not always be entirely up-to-date, but they give a good impression of the work at our department. In addition, you can find the contact details of supervisors that you can send a message. Also if you have your own project proposal matching our research scope, please don’t hesitate to contact us.