MR Perfusion

Image Slice Voxel Fit

A single dynamic from a perfusion sequence (left) and a voxel flow course with fitted Gamma variant
We will demonstrate a technique for the analysis of dynamic susceptibility contrast enhanced MR perfusion images based on Gamma Variate fitting. This combines
  • single pixel estimation of T0 values prior to fitting,
  • flexible model based principle component noise reduction,[2]
  • modified gamma variant curve fitting including T0 as a fitting parameter, [1].
Measures of T0 and TTP from the parametric images have been compared with values from regions of interest on unfitted data. The quality of the parametric image maps of T0 and TTP was assessed by two experienced neuroradiologists in normal volunteers and in 40 patients with vascular or neoplastic brain disease. Increases in SFE not correlated to SNR is seen in tumors and may provide an indication of abnormal micro-vascular structure. T0 and TTP measurements on unprocessed data demonstrated differences between gray and white matter ranging from 2.5-5.4s. Parametric images using conventional analysis techniques fail to demonstrate these differences in T0.

Library toolkit
binary 1
Sun Sparc - Solaris 2
binary 2
Intel x86 - Linux 2
[1] A.Jackson, K.L.Li, X.P.Zhu and N.A.Thacker, `Improving Time of Arrival Map Quality in MR Perfusion.' proc. MIUA, Leeds, 13-17, 1998.
[2] Zhu, X, Hutchinson C, Hawnaur J, Cootes T, Taylor C, Isherwood I. B.J.R. , 67, 976-982, 1994.
the end
[Back to software]