Embedded Files
Ex-vivo histology is the gold standard to investigate human brain microstructure. However, its invasive nature precludes its use in monitoring disease progression and the investigation of the pathophysiological origin of neurological disorders. MRStain will address this shortcoming by exploiting the sensitivity of the Magnetic Resonance Imaging (MRI) signal to estimate aggregated histological metrics in the human brain non-invasively (Fig. 1). Like established histology staining methods (e.g. myelin-basic protein), MRStain will be sensitive to changes in cellular populations, axons, myelin, and iron. This will be achieved by augmenting the MRI measurements with computational biophysical models, which can disentangle tissue metrics at the micron-scale using the macroscopic spatial resolution (1–4 mm) of MRI. However, the clinical use of these models is challenging because their validity and generalizability across disease trajectories has yet to be tested against the ex-vivo histological gold standard.
Figure 1: Relation between seven quantitative in-vivo MRI and contrast-giving microstructure visualized by blue, green, and yellow lines. The depicted MRI contrasts include images from: diffusion MRI, relaxometry, magnetization transfer imaging and quantitative susceptibility mapping. These are sensitive to three main, contrast giving microstructure properties in the brain related to the axonal/cell, myelin, and iron compartments. From our review on “Micro-structural imaging of human neocortex in vivo” (Edwards et al., 2019).
Figure 2: Pilot data illustrating proposed multi-modal dataset. A. (top to bottom): Quantitative transverse relaxation rate map (R1) overlaid with contours (red) of the excised temporal lobe (TL) and photo of the excised TL in oxygenated solution before MRI (cutting plane in yellow). B. (top to bottom): Magnification of TL region in R1 map, and MRI of the excised tissue from the same TL region about 30min. after surgery. C. (left to right): Large-scale light-microscopy and machine-learning-based quantification of axon radii in a human corpus callosum (current project using another specimen). Neurosurgery was performed at the local Department of Neurosurgery by Dr. Thomas Sauvigny with whom I will collaborate on the recruitment of drug-resistant TL epilepsy patients throughout the proposed project (LOC_Saugivny).
This project will address this shortcoming using the drug-resistant temporal lobe epilepsy disease to generate a globally unique multi-modal dataset that combines novel in-vivo and ex-vivo MRI techniques with biophysical models as well as cutting-edge large-scale 3D histology (Fig. 2). The project will benefit from a unique translational university hospital environment where large sections of freshly excised brain tissue from drug-resistant temporal lobe epilepsy patients (TLE, 80 sections of about 3 x 2 x cm^3) can be examined, enabling in-vivo MRI-based biophysical tissue parameters to be validated against their histological gold standard. We will develop an MRStain model to identify TLE relevant changes that will achieve a paradigm shift in how epilepsy is treated by identifying target brain areas for surgery that will help to predict seizure-free outcome after surgery. The final validated MRStain models will also pave the way for similar noninvasive investigations of other neurological and neuropsychiatric diseases with unprecedented precision.
As part of the imaging work package of MRStain, it is intended to acquire a unique multi-modal dataset from up to 80 TLE patients including in-vivo MRI, ex-vivo MRI of fresh and unfixed tissue, ultra-high-resolution ex-vivo MRI of fixed tissue, and 3D ex-vivo histology. The histology data will be segmented and quantitative micro-scale parameters (e.g., cell density) will be extracted and stacked into 3D maps. The MRI and histology data of the same persons will be spatially integrated such that they can be used use to train and validate in vivo histology computational models of tissue microstructure using in vivo MRI.
Page updated
Google Sites
Report abuse