Research Projects of the Division of Medical and Biological Informatics
Diagnostic support in neurodegenerative diseases
Visualization of atrophic alterations in the hippocampal region of the brain, as determined by using the atlas-based measuring technique developed here. Color scale ranges from blue to red for mild to severe atrophy. A: Healthy subject, B: Alzheimer patient
| © dkfz.de
Magnetic resonance imaging (MRI) and diffusion-weighted MRI (dMRI) facilitate non-invasive anatomical and microstructural measurements on the living brain. A reliable use of these techniques - for example to detect Alzheimer's disease at as early a stage as possible - is presently hindered by several central challenges.
We focus on atlas-based measuring methods for macroscopic atrophy by using MRI as well as the measurement of microscopic fiber integrity by using dMRI. Studies conducted on persons with mild cognitive impairment have shown that brain damage already occurs at a much earlier time and can be detected sooner than the diagnosis with conventional clinical procedures allows.
Development and evaluation of new imaging techniques
Above: Volume visualization of two different 90° filament yarnmodels. Below: Visualization of voxel-based diffusion models in a 90° crossing (a) and a 45° crossing (b) of filament yarnmodels
| © dkfz.de Artificial filament yarn models of neural tracts with properties as realistic as possible are being developed to facilitate a controlled examination of quantitative measuring methods under different conditions and especially in fiber crossings and in partial volumes.
Reconstruction of neural tracts in the brain
Visualization of reconstructed neural tracts
| © dkfz.de The analysis of diffusion-weighted image volumes makes it possible to identify and quantitatively evaluate damages to the complex neuronal network of the brain that consists of individual nerve fibers. However, many of the relevant nerve bundles are very thin, significantly bent, lie densely packed together or cross one another. They are partially also only compromised at certain locations, making a complete identification difficult without knowing the normal course. Because of the difference in the anatomical structures and their interconnectedness, there is still no procedure available to comprehensively analyze all relevant nerve bundles in a robust and reproducible manner. For example, a central challenge is suitably modeling the properties and special characteristics of the relevant structures and to include this as a priori knowledge when considering specific structures.
Voxel-based group statistics of microscopic alterations
The reduction of the group average neural tracts to their midline and a corresponding accumulation of the measurements along the direction of reduction facilitates group comparisons and the extraction of disease-specific injury patterns.
| © dkfz.de In medicine detailed information about neuronal networking patterns is an important basis for comprehensively understanding the brain and its diseases. In this regard dMRI facilitates voxel-based measurements in living human beings; the analysis of this information is especially interesting if it can be assigned to specific structures and then be statistically analyzed over different patients or subject groups.
MITK diffusion imaging
Screen shot of the MITK-DI module as an application component in MITK 3M3
| © dkfz.de In order to further support development and sustainable evaluation in DTI and HARDI research, we have developed and publicly provided software tools for data I/O, model reconstruction, interactive visualization, feature extraction, statistics and other diffusion image processing tasks. The software is published as a toolkit component of the Medical Imaging and Interaction Toolkit (MITK, www.mitk.org).
Selected publications
- [J03-10] Fritzsche KH, Stieltjes B, Schlindwein S, van Bruggen T, Essig M, Meinzer HP. Automated MR morphometry to predict Alzheimer's disease in mild cognitive impairment. Int J Comput Assist Radiol Surg 2010 May 4. [Epub ahead of print] dx.doi.org/10.1007/s11548-010-0412-0.
- [J01-10] Fritzsche KH, Laun FB, Meinzer HP, Stieltjes B. Opportunities and pitfalls in the quantification of fiber integrity: What can we gain from Q-ball imaging? Neuroimage 2010 May;51(1):242-51. 10.1016/j.neuroimage.2010.02.007.
- [P17-10] Moussavi A, Stieltjes B, Fritzsche KH, Laun FB. Novel spherical phantoms for Q-ball imaging under in-vivo conditions, In Proceedings 19th Scientific Meeting, International Society for Magnetic Resonance in Medicine, Stockholm, Sweden, 2010.
- [P16-10] Fritzsche K, Meinzer H-P. MITK-DI -
- [J02-08] Fritzsche KH, Wangenheim A, Abdala DD, Meinzer HP. A computational method for the estimation of atrophic changes in Alzheimers disease and mild cognitive impairment, Comput Med Imaging Graph 32 (2008) 294-303.
