Evaluation of brain diseases by quantitative magnetic resonance metabolic, perfusion and diffusion imaging

Exploitable results

Three different MR techniques have been included in the project, i.e. magnetic resonance spectroscopic imaging (MRSI), water diffusion imaging, and perfusion imaging. Spectroscopic Imaging The first purpose of this task was to develop a single-slice metabolic imaging pulse sequence, delivering quantitatively metabolite concentrations. A crucial point was that the centres participating in the clinical part of the project should be able to implement the pulse sequence on their respective MR scanners. This has lead to the choice of a multi-echo sequence with one spatial phase encoding step per repetition and a PRESS pre selection to reduce lipid signals from the scalp. Choline compounds, total creatine, N-acetylaspartate and lactate will be studied. For multi-centre studies the chosen sequence has the additional advantage that institutes with no capabilities of performing multi-echo experiments can apply two or more single echo sequences with different echo times. Preliminary in vivo studies with the sequence were carried out on rats. In order to obtain a �golden standard� for comparison with values obtained by the developed MRSI protocol, measurements of T1 and T2 of water and metabolites as well as calculations of metabolite concentrations were performed according to a single voxel protocol developed in a previous study. Results were obtained by four different centres examining six normal volunteers each. Adult brain-water relaxation times appear to be substantially shorter than neonatal values. The underlying reasons for intergroup systematic errors are probably related to subtle differences in data-acquisition and -analysis methods. However, results compared favourably with those reported in previous studies and it is appropriate to accept these values as �golden standards� for purposes of comparison and interpretation. The effect of an initial spoiling pulse to define the relaxation delay was also investigated and it was found to have no effect on T2 measurements but T1 was underestimated 10 to 20 % when the pulse was absent. Cohort mean results obtained by single voxel spectroscopy and MRSI, respectively, were compared, and it was found that T1 and T2 for brain water compared very well. Also Cho T2s and concentrations as well as NAA concentrations were in good agreement, but some deviations for NAA and Cr T2 as well as for Cr concentrations were found. To clear up this matter, studies are carried out at the involved centres. A multi-slice, multi-echo MRSI sequence was also developed and tested on phantoms and subsequently on Wistar rats with injected tumour cells. Images where successfully acquired in two contiguous slices with two echoes. It was concluded that the multi-slice, multi-echo sequence could be successfully implemented for animal models. In order to reduce the measuring time even further on the clinical whole body scanner an echo-planar spectroscopy (EPS) sequence was implemented here. EPS has some advantages over conventional MRSI. It offers flexible trade-offs between signal/noise ratio and measurement time, with very reduced time required for obtaining the water reference signal. The sequence is conceptually similar to implementations by other groups, but applies low acquisition bandwidth and non-interleaved sampling with the advantage of low hardware requirements and reduced acquisition time. To have an estimate of localisation performance a simple test was needed. It was demonstrated how a narrow tube with water and an inner diameter up to half the pixel width can be used to determine the localisation performance and signal loss in the multi-slice experiments. A dedicated software package for post processing of MRSI data was also developed. Diffusion Imaging It was decided that for the clinical studies a multi-slice EPI sequence should be used if possible. To eliminate the effect of anisotropy the trace of the diffusion tensor should be obtained by taking the sum of three perpendicular measurements, applying two b-values in each direction. Additionally a diffusion insensitive measurement should be performed. Phantom experiments were performed on pure water, acetone, and 2-propanol to evaluate the sequences. Some validation studies are still being carried out. Perfusion Imaging The first task was to develop a robust method for assessing cerebral perfusion and blood volume in a minimal invasive way, using bolus injections of the contrast agent Gd-DTPA during fast imaging. A blipped EPI sequence was implemented using a matrix of 128 x 128 and an echo time of 29 ms. Thus the bolus passage can be followed with 6 anatomical slices every second with a reasonably good signal/noise ratio. The relationship between the concentration of Gd-DTPA and MR variables was studied. Animal studies at 4.7 T showed higher relative signal changes at low echo times than at long echo times, corresponding to a exp(TE0.9 R2*) dependence. This shows the importance of all centres using the same echo time in the clinical investigations. No deviations from the linear relation between dR2* and Gd concentration could however be detected at concentrations up to 2-3 mM. A total of 11 young healthy volunteers were investigated by PET scanning and subsequently by MR scanning. Both examinations were performed during rest with inhalation of atmospheric air as well as during voluntary hyperventilation and during inhalation of atmospheric air containing 6% CO2. The values found for CBF and CBV by deconvolution analysis of MR data exceeded those found by PET scanning by a factor of 3-4. The changes during the experimental conditions, however, were qualitatively in accordance with each other. The discrepancies may be due to different relations between the Gd concentration and change in dR2* for the arterial and the tissue signal. It was also investigated whether spin labelling methods can be used to determine perfusion in the brain with sufficient accuracy. Contrary to the invasive method discussed above, the spin labelling methods use endogenous water as tracer.

Suche nach OpenAIRE-Daten ...

Exploitable results

Diese Seite teilen

Herunterladen