tina [about] [news] [software] [demos] [projects] [docs] [images] [faq] [people] [teaching] [tina4] [tina5] [WIKI] [KNOPPIX

project : Net Cerebral Blood Flow

jump to



The aim of this work is to provide an alternative method to the standard perfusion techniques for determining cerebral blood flow from images obtained using dynamic susceptibility contrast-enhanced magnetic resonance imaging (DSCE-MRI).
A bolus of the paramagnetic contrast agent Gd-DTPA through the brain results in a drop in signal intensity, the log change is proportional to contrast agent concentration. The area under the graph of the contrast agent concentration over time is therefore proportional to the volume of contrast agent which has passed through the voxel (CBV). The mean transit time (MTT) of the contrast from the arterial to venous pools is standardly computed using a deconvolution of the voxel residue function (ie, the amount of contrast still remaining in the voxel at a certain time) by the Arterial Input Function (AIF) (ie, the contrast-time curve measured in an artery). Cerebral blood flow (relative) is then calculated as CBV/MTT.
This technique suffers several drawbacks, most notably that the mean transit time cannot be reliably calculated.
We have developed a novel technique to obtain a Net Mean Transit Time and hence absolute Net Cerebral Blood Flow for every imaged voxel. Instead of using deconvolution to obtain an MTT, we calculate the average time of arrival of the contrast in each voxel and use spatial differentiation of these times to calculate the Net mean transit time of contrast across the voxel. CBV can be made absolute by normalising to a voxel which is 100% blood (in an artery or vein). Net cerebral blood flow is therefore CBV/NMTT scaled by a factor to transform the value into the standard physiological units of ml/100g/min.

Developers Marietta Scott email: marietta.scott@man.ac.uk
Neil Thacker email: neil.thacker@man.ac.uk


This work was funded as part of a Wellcome Trust project on relating cross-sectional and longitudinal changes in brain function to cognitive function in normal old age.


For a demonstration of the software associated with this project, please download Tina5 and build the NCBF example toolkit under tina5/tina-tools/toolkits/ncbf/.


Marietta's PhD thesis entitled ``Towards a Quantitative Methodology for the Assessment of Cerebral Blood Flow in Magnetic Resonance Imaging'', published Summer 2005 by the University of Manchester is available below as a pdf file and as text:
Thesis as a pdf

Thesis as text

Valid HTML 4.01!