Can dynamic susceptibility contrast magnetic resonance imaging perfusion data be analyzed using a model based on directional flow?

PURPOSE

To examine the implications of a physiological model of cerebral blood that uses the contradictory assumption that blood flow in all voxels of DSCE-MRI data sets is directional in nature. Analysis of dynamic susceptibility contrast-enhanced magnetic resonance imaging (DSCE-MRI) uses techniques based on indicator dilution theory. Underlying this approach is an assumption that blood flow through pixels of gray and white matter is entirely random in direction.

MATERIALS AND METHODS

We have used a directional flow model to estimate theoretical blood flow velocities that would be observed through normal cerebral tissues. Estimates of flow velocities from individual pixels were made by measuring the mean transit time for net flow (nMTT). Measurements of nMTT were made for each voxel by estimating the mean difference in contrast arrival time between each of the adjacent six voxels.

RESULTS

Examination of the spatial distribution of contrast arrival time from DSCE-MRI data sets in normal volunteers demonstrated clear evidence of directional flow both in large vessels and in gray and white matter. The mean velocities of blood flow in gray and white matter in 12 normal volunteers were 0.25 +/- 0.013 and 0.21 +/- 0.014 cm/second, respectively, compared to predicted values of 0.25 and 0.18 cm/second. These values give measured nMTT for a 1-mm isotropic voxel of gray and white matter of 0.45 +/- 0.12 and 0.52 +/- 0.11 seconds, respectively, compared to predicted values of 0.47 and 0.55 seconds.

CONCLUSION

A directional model of blood flow provides an alternative approach to the calculation of cerebral blood flow from (CBF) DSCE-MRI data.