Boudewijn Lelieveldt

‘The mix of competencies in the Medical Delta alliance makes it possible to establish unique collaborations between engineers, clinicians and neuroscientists.’



Boudewijn Lelieveldt earned his PhD in 1999 and then became university lecturer and associate professor at the Laboratory for Clinical and Experimental Image Processing at LUMC. He has been Professor of Biomedical Imaging at the department of Radiology at LUMC and at the Intelligent Systems group at the Faculty Electrical Engineering, Mathematics and Computer Science at TU Delft since 2011.


TU Delft: Multimedia signal processing
LUMC: Biomedical Imaging at Radiology

The medical knowledge broker

MRI images of the blood flow in the heart

‘What I like to do is to look for inter- esting niche imaging markets. By combining our expertise in this field in Leiden with the complementary expertise in Delft. Let me give an example: MRI images of the blood flow in the heart. Nowadays, it is possible to measure the velocity of every image pixel. Have a look... here at the top you can see normal MRI images of the heart, which give a cardiologist quite a bit of informa- tion. But in these pictures down here at the bottom, the gray values give the velocities with which the blood flows through the heart– from the left, to the right, to the front and to the back.’

‘In addition to the expertise required to produce the images, we also have the software to measure the velocities, and thanks to our col- laboration with Delft we are also able to visualize this complex infor- mation. What you see here is still to a large extent our own work, but thanks to our connection with Delft we now have great insight into all the data. This will enable us to really make something that no one else in the world has.’

Atlas of the brain based on gene expression

‘Another example of the collabora- tion with Delft – and also another amazing piece of engineering. The Allen Institute for Brain Science – Paul Allen is one of the founders of Microsoft – has, over the past ten years, developed into an institute of international repute. And particu- larly the complexity of their data is enormous. They have made an atlas of the brain, but one based on gene expression. In which part of the brain are which genes switched ‘on’? Considering there are more than 20,000 genes, this is an enormously complex dataset. How should you ‘look’ at this? How do you obtain information from this? Thanks to the new concept of ‘spatial gene expression’ one can see the data as an image process- ing problem. And that is our bread and butter.’

Analyze higher dimensional datasets

‘Together with bio-informaticians from Delft we delved into this five years ago. I talked to a colleague at TU Delft about the type of data in the Allen Brain database. This col- league works on image recognition and is an expert in the field of image tracking, and he mentioned that he had developed a method to analyze these kinds of higher dimensional datasets (in which every point repre- sents not two or three but thousands of values – like gene expression) during his PhD research.’

‘This method was not affected by the usual problems experienced when analyzing this type of data. My first thought was: I have to see it before I believe it. But then I came across an increasing number of articles in various leading journals in which his method had been used. So it really was an excellent PhD paper! And exactly what we needed! And I’d been sitting close to the author all these years! This is a great example of the uniqueness of the Medical Delta alliance: the mix of competencies, close to each other, makes it possible to establish unique collaborations.’

Visualize pixels and their gene expression pattern

‘The underlying hypothesis is that genes with a similar gene expression pattern may also form a functional unit. Thanks to my col- league from Delft we can visualize which pixels have the same gene expression pattern and which ones do something totally different. If you adapt the information to the anatomical structure, then you see that the division corresponds very well with the anatomical data. This is great proof that we are on the right track!’

‘The Allen Institute very recently published two papers on a study in which the expression of a subset of genes in the brain of autistic children was investigated. It showed that the layered structure (as seen in gene expression) of the frontal cor- tex in these children really is very dif- ferent. With regard to our research here, we will focus on clever ways of extracting as much information as possible from this data, and apply- ing it to brain research together with clinicians and neuroscientists.’