One week in the clinic - by Masoomeh Rahimpour
Within HYBRID, each of the fellows spends time in the medical imaging departments of a hospital, to familiarize themselves with clinical workflows. Masoomeh Rahimpour spent one week at the MRI and PET centres of UZ Leuven. She learned about the various imaging techniques that are used in the diagnosis and treatment planning of patients with brain tumours.
Different imaging modalities show varying information about the tissue type and underlying disease. For example, structural MRI provides detailed information on the structure of the brain to facilitate the segmentation of brain tumours. Segmentation is the ‘drawing’ of a certain tissue in the MRI image. The goal of brain tumour segmentation is to detect the location and extension of the tumoral regions. This is a crucial step for diagnosis and treatment planning, but it does not provide the surgeons with all the information they need to plan a brain surgery procedure. The patient might need to undergo other types of MRI imaging such as Functional MRI, and Diffusion Tensor Imaging to get more information about the surroundings of the tumour and preserve brain function during surgery. Furthermore, some types of tumours are not easily detectable with MRI and another imaging modality, called Positron Emission Tomography (PET), is needed to better differentiate them.
Following I will briefly explain what I have learnt about these imaging modalities from my visit to the MRI and PET centres at UZ Leuven.
Functional MRI (fMRI) allows us to learn which part of the brain is being employed when performing a particular task. When we are involved in different tasks, the activity of the neurons in the brain continually changes. These changes in activity can be measured by fMRI. The basic idea is that when a neuron becomes active, it needs more blood; this is a signal which is picked up by fMRI.
What makes fMRI an important tool for preoperative planning in brain tumour surgery is that it can show the surgeon where the important structures in the brain are. For example areas that control movement, speech, and memory. To localize these brain functions, patients are asked to perform some tasks during an fMRI. For example, to precisely image the locations in the brain that control movement, patients are asked to tap with their fingers and toes.
Diffusion Tensor Imaging
Diffusion Tensor Imaging is a type of MRI that can identify connections in the brain. It detects how water in the brain travels along the white matter tracts. The white matter tracts connect different parts of the brain and must be protected during the surgery. DTI images are loaded into the navigation systems that are used in the operating room to serve as a GPS and map for the surgeons. This way they can remove tumours to the greatest extent possible without harming areas of the brain that are critical to a patient’s quality of life.
Positron Emission Tomography
MRI techniques cannot reveal the abnormalities in the cellular level. For that you need Positron Emission Tomography (PET). PET measures the activity of the cells of body tissues by using a radioactive tracer (radiotracer). Radiotracers are molecules that are very similar to molecules that are used by the cells in our body. One of the most-commonly used radiotracers in brain imaging looks a lot like glucose, since brain cells use glucose. Radiotracers have a special characteristic: they are radioactively labelled*, so that the PET-scanner can detect them. Abnormal tissues such as tumours in the brain appear brighter in the PET scans because they use more glucose than normal tissues.
(*PET tracers are injected in so small amounts that the radioactivity does not harm the patient.)
PET discovers differences in tumour regions, which might not be visible on an MRI scan. Some parts of the tumour use more glucose than other parts, meaning the tumour is ‘metabolically active’. Among others, PET can be used to assess the effect of treatment. It can show changes in the tumour earlier than MRI. The tumour might already be less metabolically active, while it still has the same size.
Although each imaging modality contains valuable information about the type of disease and patient condition, it is not always straightforward to extract the relevant information. The quality of image might degrade due to the patient motion, incorrect scanner setting, and limited scan time per patient. Different protocols to obtain the images in different imaging centres also lead to heterogeneous datasets where some of the acquired images are missed or are unusable. Being aware of the practical challenges which might occur during the clinical imaging allows us to be more critical in developing and implementation of effective ways to process the images from different modalities. We need to think of more generic and flexible methods, which are able to comply with the clinically relevant scenarios.
Image from https://www.uzleuven.be/en